diff --git "a/notebooks/lilabc_test-island.ipynb" "b/notebooks/lilabc_test-island.ipynb"
new file mode 100644--- /dev/null
+++ "b/notebooks/lilabc_test-island.ipynb"
@@ -0,0 +1,2706 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import pandas as pd\n",
+    "import seaborn as sns\n",
+    "\n",
+    "sns.set_style(\"whitegrid\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Reading in our potential test sets from [here](https://huggingface.co/datasets/imageomics/lila-bc-camera/tree/d18307b285217d18b31d1a7b2c9091bb0873ade0/data/potential-test-sets):\n",
+    "  - [Ohio Small Animals](https://lila.science/datasets/ohio-small-animals/)\n",
+    "  - [Desert Lion Conservation Camera Traps](https://lila.science/datasets/desert-lion-conservation-camera-traps/)\n",
+    "  - [Orinoquia Camera Traps](https://lila.science/datasets/orinoquia-camera-traps/)\n",
+    "  - [Island Conservation Camera Traps](https://lila.science/datasets/island-conservation-camera-traps/)\n",
+    "  - [ENA24](https://lila.science/datasets/ena24detection)\n",
+    "\n",
+    "We'll clean them down to just the taxa and identifier columns, then further reduce to make balanced test sets for each.\n",
+    "\n",
+    "# Island Conservation Camera Traps Datasets\n",
+    "upper/lower bounded and balanced"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>dataset_name</th>\n",
+       "      <th>url_gcp</th>\n",
+       "      <th>url_aws</th>\n",
+       "      <th>url_azure</th>\n",
+       "      <th>image_id</th>\n",
+       "      <th>sequence_id</th>\n",
+       "      <th>location_id</th>\n",
+       "      <th>frame_num</th>\n",
+       "      <th>original_label</th>\n",
+       "      <th>scientific_name</th>\n",
+       "      <th>...</th>\n",
+       "      <th>superfamily</th>\n",
+       "      <th>family</th>\n",
+       "      <th>subfamily</th>\n",
+       "      <th>tribe</th>\n",
+       "      <th>genus</th>\n",
+       "      <th>species</th>\n",
+       "      <th>subspecies</th>\n",
+       "      <th>variety</th>\n",
+       "      <th>multi_species</th>\n",
+       "      <th>num_species</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>cow</td>\n",
+       "      <td>bos taurus</td>\n",
+       "      <td>...</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>bovidae</td>\n",
+       "      <td>bovinae</td>\n",
+       "      <td>bovini</td>\n",
+       "      <td>bos</td>\n",
+       "      <td>bos taurus</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>False</td>\n",
+       "      <td>1.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>donkey</td>\n",
+       "      <td>equus asinus</td>\n",
+       "      <td>...</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>equidae</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>equus</td>\n",
+       "      <td>equus asinus</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>False</td>\n",
+       "      <td>1.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>donkey</td>\n",
+       "      <td>equus asinus</td>\n",
+       "      <td>...</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>equidae</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>equus</td>\n",
+       "      <td>equus asinus</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>False</td>\n",
+       "      <td>1.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>...</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>iguanidae</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>False</td>\n",
+       "      <td>1.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>...</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>iguanidae</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>False</td>\n",
+       "      <td>1.0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>5 rows × 34 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                       dataset_name  \\\n",
+       "0  Island Conservation Camera Traps   \n",
+       "1  Island Conservation Camera Traps   \n",
+       "2  Island Conservation Camera Traps   \n",
+       "3  Island Conservation Camera Traps   \n",
+       "4  Island Conservation Camera Traps   \n",
+       "\n",
+       "                                             url_gcp  \\\n",
+       "0  https://storage.googleapis.com/public-datasets...   \n",
+       "1  https://storage.googleapis.com/public-datasets...   \n",
+       "2  https://storage.googleapis.com/public-datasets...   \n",
+       "3  https://storage.googleapis.com/public-datasets...   \n",
+       "4  https://storage.googleapis.com/public-datasets...   \n",
+       "\n",
+       "                                             url_aws  \\\n",
+       "0  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "1  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "2  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "3  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "4  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "\n",
+       "                                           url_azure  \\\n",
+       "0  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "1  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "2  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "3  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "4  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "\n",
+       "                                            image_id  \\\n",
+       "0  Island Conservation Camera Traps : dominicanre...   \n",
+       "1  Island Conservation Camera Traps : dominicanre...   \n",
+       "2  Island Conservation Camera Traps : dominicanre...   \n",
+       "3  Island Conservation Camera Traps : dominicanre...   \n",
+       "4  Island Conservation Camera Traps : dominicanre...   \n",
+       "\n",
+       "                                  sequence_id  \\\n",
+       "0  Island Conservation Camera Traps : unknown   \n",
+       "1  Island Conservation Camera Traps : unknown   \n",
+       "2  Island Conservation Camera Traps : unknown   \n",
+       "3  Island Conservation Camera Traps : unknown   \n",
+       "4  Island Conservation Camera Traps : unknown   \n",
+       "\n",
+       "                                         location_id  frame_num  \\\n",
+       "0  Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "1  Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "2  Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "3  Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "4  Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "\n",
+       "  original_label scientific_name  ... superfamily     family subfamily  \\\n",
+       "0            cow      bos taurus  ...         NaN    bovidae   bovinae   \n",
+       "1         donkey    equus asinus  ...         NaN    equidae       NaN   \n",
+       "2         donkey    equus asinus  ...         NaN    equidae       NaN   \n",
+       "3         iguana          iguana  ...         NaN  iguanidae       NaN   \n",
+       "4         iguana          iguana  ...         NaN  iguanidae       NaN   \n",
+       "\n",
+       "    tribe   genus       species subspecies variety multi_species num_species  \n",
+       "0  bovini     bos    bos taurus        NaN     NaN         False         1.0  \n",
+       "1     NaN   equus  equus asinus        NaN     NaN         False         1.0  \n",
+       "2     NaN   equus  equus asinus        NaN     NaN         False         1.0  \n",
+       "3     NaN  iguana           NaN        NaN     NaN         False         1.0  \n",
+       "4     NaN  iguana           NaN        NaN     NaN         False         1.0  \n",
+       "\n",
+       "[5 rows x 34 columns]"
+      ]
+     },
+     "execution_count": 2,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df = pd.read_csv(\"../data/potential-test-sets/Island_Conservation_Camera_Traps_image_urls_and_labels.csv\", low_memory = False)\n",
+    "df.head()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "Index(['dataset_name', 'url_gcp', 'url_aws', 'url_azure', 'image_id',\n",
+       "       'sequence_id', 'location_id', 'frame_num', 'original_label',\n",
+       "       'scientific_name', 'common_name', 'datetime', 'annotation_level',\n",
+       "       'kingdom', 'phylum', 'subphylum', 'superclass', 'class', 'subclass',\n",
+       "       'infraclass', 'superorder', 'order', 'suborder', 'infraorder',\n",
+       "       'superfamily', 'family', 'subfamily', 'tribe', 'genus', 'species',\n",
+       "       'subspecies', 'variety', 'multi_species', 'num_species'],\n",
+       "      dtype='object')"
+      ]
+     },
+     "execution_count": 3,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df.columns"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Observe that we also now get multiple URL options; `url_aws` will likely be best/fastest for use with [`distributed-downloader`](https://github.com/Imageomics/distributed-downloader) to get the images."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "<class 'pandas.core.frame.DataFrame'>\n",
+      "RangeIndex: 44438 entries, 0 to 44437\n",
+      "Data columns (total 34 columns):\n",
+      " #   Column            Non-Null Count  Dtype  \n",
+      "---  ------            --------------  -----  \n",
+      " 0   dataset_name      44438 non-null  object \n",
+      " 1   url_gcp           44438 non-null  object \n",
+      " 2   url_aws           44438 non-null  object \n",
+      " 3   url_azure         44438 non-null  object \n",
+      " 4   image_id          44438 non-null  object \n",
+      " 5   sequence_id       44438 non-null  object \n",
+      " 6   location_id       44438 non-null  object \n",
+      " 7   frame_num         44438 non-null  int64  \n",
+      " 8   original_label    44438 non-null  object \n",
+      " 9   scientific_name   44438 non-null  object \n",
+      " 10  common_name       44438 non-null  object \n",
+      " 11  datetime          44354 non-null  object \n",
+      " 12  annotation_level  44438 non-null  object \n",
+      " 13  kingdom           44438 non-null  object \n",
+      " 14  phylum            44438 non-null  object \n",
+      " 15  subphylum         44438 non-null  object \n",
+      " 16  superclass        79 non-null     object \n",
+      " 17  class             44438 non-null  object \n",
+      " 18  subclass          25269 non-null  object \n",
+      " 19  infraclass        25189 non-null  object \n",
+      " 20  superorder        25268 non-null  object \n",
+      " 21  order             44261 non-null  object \n",
+      " 22  suborder          18430 non-null  object \n",
+      " 23  infraorder        79 non-null     object \n",
+      " 24  superfamily       6945 non-null   object \n",
+      " 25  family            44190 non-null  object \n",
+      " 26  subfamily         16736 non-null  object \n",
+      " 27  tribe             11810 non-null  object \n",
+      " 28  genus             28006 non-null  object \n",
+      " 29  species           13202 non-null  object \n",
+      " 30  subspecies        0 non-null      float64\n",
+      " 31  variety           938 non-null    object \n",
+      " 32  multi_species     44438 non-null  bool   \n",
+      " 33  num_species       44438 non-null  float64\n",
+      "dtypes: bool(1), float64(2), int64(1), object(30)\n",
+      "memory usage: 11.2+ MB\n"
+     ]
+    }
+   ],
+   "source": [
+    "df.info(show_counts = True)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Most of these are not labeled to the species level, so we'll need to take a closer look at the taxa. Perhaps subdivide?\n",
+    "\n",
+    "Everything is labeled to the class level, most make it to family, but then it's a steep drop-off."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "lin_taxa = ['kingdom', 'phylum', 'class', 'order', 'family', 'genus', 'species']\n",
+    "taxa_cols = ['original_label', 'scientific_name', 'common_name', 'kingdom', 'phylum', 'class', 'order', 'family', 'genus', 'species']"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "original_label     45\n",
+       "scientific_name    42\n",
+       "common_name        45\n",
+       "kingdom             1\n",
+       "phylum              2\n",
+       "class               6\n",
+       "order              19\n",
+       "family             25\n",
+       "genus              25\n",
+       "species            22\n",
+       "dtype: int64"
+      ]
+     },
+     "execution_count": 6,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df[taxa_cols].nunique()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "42 scientific names, but 45 common and original. Only 22 species --more genera and families than species!\n",
+    "\n",
+    "Also, should check for duplicated images."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "number of unique images: 44007\n"
+     ]
+    },
+    {
+     "data": {
+      "text/plain": [
+       "multi_species\n",
+       "False    43577\n",
+       "True       861\n",
+       "Name: count, dtype: int64"
+      ]
+     },
+     "execution_count": 7,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "print(f\"number of unique images: {df[\"image_id\"].nunique()}\")\n",
+    "\n",
+    "df[\"multi_species\"].value_counts()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "We also have 430 duplicated images (multiple species per image)."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "num_species\n",
+       "1.0    43577\n",
+       "2.0      858\n",
+       "3.0        3\n",
+       "Name: count, dtype: int64"
+      ]
+     },
+     "execution_count": 8,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df[\"num_species\"].value_counts()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "scientific_name\n",
+       "gallus gallus domesticus    308\n",
+       "procellariidae              180\n",
+       "asio flammeus                85\n",
+       "aves                         83\n",
+       "rattus                       36\n",
+       "felis catus                  29\n",
+       "leporidae                    27\n",
+       "iguana                       24\n",
+       "calonectris                  23\n",
+       "calcinus tubularis           20\n",
+       "corvus corax                 13\n",
+       "capra hircus                  8\n",
+       "insecta                       5\n",
+       "brachyura                     4\n",
+       "falco tinnunculus             3\n",
+       "sus scrofa                    3\n",
+       "equus asinus                  3\n",
+       "columbidae                    2\n",
+       "zenaida asiatica              1\n",
+       "bos taurus                    1\n",
+       "rallidae                      1\n",
+       "lepidoptera                   1\n",
+       "araneae                       1\n",
+       "Name: count, dtype: int64"
+      ]
+     },
+     "execution_count": 9,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df.loc[df[\"multi_species\"], \"scientific_name\"].value_counts()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "That's a large variety, what happens if we drop these?"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "41"
+      ]
+     },
+     "execution_count": 10,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df.loc[~df[\"multi_species\"], \"scientific_name\"].nunique()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Hmm, we'll lose one."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 11,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "['bos taurus' 'equus asinus' 'iguana' 'corvus corax' 'felis catus'\n",
+      " 'canis familiaris' 'rattus' 'zenaida asiatica' 'aves' 'passeriformes'\n",
+      " 'strigiformes' 'athene cunicularia' 'strix varia' 'butorides virescens'\n",
+      " 'falco sparverius' 'mimidae' 'ardea herodias' 'gallus gallus domesticus'\n",
+      " 'procellariidae' 'leporidae' 'falco tinnunculus' 'calonectris'\n",
+      " 'nasua nasua' 'hydrobates pelagicus' 'columbidae' 'araneae'\n",
+      " 'asio flammeus' 'insecta' 'turdus' 'phocidae' 'capra hircus' 'sus scrofa'\n",
+      " 'nyctanassa violacea' 'chelonioidea' 'brachyura' 'megapodiidae'\n",
+      " 'anous stolidus' 'caloenas nicobarica' 'rallidae' 'varanidae'\n",
+      " 'lepidoptera' 'calcinus tubularis'] ['cow' 'donkey' 'iguana' 'raven' 'cat' 'dog' 'rat' 'white-winged_dove'\n",
+      " 'bird' 'passerine' 'owl' 'burrowing_owl' 'barred_owl' 'green_heron'\n",
+      " 'american_kestrel' 'mockingbird' 'great_blue_heron' 'chicken' 'petrel'\n",
+      " 'rabbit' 'petrel_chick' 'kestrel' 'shearwater' 'coati' 'storm_petrel'\n",
+      " 'pigeon' 'dove' 'spider' 'short-eared_owl' 'insect' 'zorzal' 'seal'\n",
+      " 'goat' 'pig' 'yellow-crowned_night_heron' 'sea_turtle' 'rooster' 'crab'\n",
+      " 'megapode' 'brown_noddy' 'nicobar_pigeon' 'rail' 'monitor_lizard' 'moth'\n",
+      " 'hermit_crab']\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(df.scientific_name.unique(), df.original_label.unique())"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "common_name\n",
+       "rabbits and hares    7729\n",
+       "rat                  6880\n",
+       "petrel               6629\n",
+       "typical iguanas      6187\n",
+       "shearwater           1722\n",
+       "petrel chick         1097\n",
+       "monitor lizard        354\n",
+       "megapode              274\n",
+       "bird                  137\n",
+       "rail                   52\n",
+       "insect                 40\n",
+       "songbirds              38\n",
+       "pigeon                 30\n",
+       "crab                   20\n",
+       "thrush                 15\n",
+       "seal                   15\n",
+       "sea turtle              6\n",
+       "spider                  5\n",
+       "mockingbird             3\n",
+       "dove                    1\n",
+       "owl                     1\n",
+       "moth                    1\n",
+       "Name: count, dtype: int64"
+      ]
+     },
+     "execution_count": 12,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df.loc[df[\"species\"].isna(), \"common_name\"].value_counts()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Perhaps this one should be a common name task since they are provided for all images.\n",
+    "\n",
+    "\"typical iguanas\" and \"rabbits and hares\" may not be ideal.\n",
+    "Maybe \"typical iguanas\" is meant to be \"iguana iguana\" like representative species? let's check it's scientific name."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 14,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "scientific_name\n",
+       "iguana    6187\n",
+       "Name: count, dtype: int64"
+      ]
+     },
+     "execution_count": 14,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df.loc[df[\"common_name\"] == \"typical iguanas\", \"scientific_name\"].value_counts()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Just iguana. How about \"rabbits and hares\"?"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 15,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "scientific_name\n",
+       "leporidae    7729\n",
+       "Name: count, dtype: int64"
+      ]
+     },
+     "execution_count": 15,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df.loc[df[\"common_name\"] == \"rabbits and hares\", \"scientific_name\"].value_counts()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "array(['cattle', 'donkey', 'typical iguanas', 'raven', 'cat',\n",
+       "       'domestic dog', 'rat', 'white-winged dove', 'bird', 'songbirds',\n",
+       "       'owl', 'burrowing owl', 'barred owl', 'green heron',\n",
+       "       'american kestrel', 'mockingbird', 'great blue heron',\n",
+       "       'domestic chicken', 'petrel', 'rabbits and hares', 'petrel chick',\n",
+       "       'kestrel', 'shearwater', 'coati', 'storm petrel', 'pigeon', 'dove',\n",
+       "       'spider', 'short-eared owl', 'insect', 'thrush', 'seal', 'goat',\n",
+       "       'pig', 'yellow-crowned night heron', 'sea turtle', 'rooster',\n",
+       "       'crab', 'megapode', 'brown noddy', 'nicobar pigeon', 'rail',\n",
+       "       'monitor lizard', 'moth', 'hermit crab'], dtype=object)"
+      ]
+     },
+     "execution_count": 13,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df.common_name.unique()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 17,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>dataset_name</th>\n",
+       "      <th>url_gcp</th>\n",
+       "      <th>url_aws</th>\n",
+       "      <th>url_azure</th>\n",
+       "      <th>image_id</th>\n",
+       "      <th>sequence_id</th>\n",
+       "      <th>location_id</th>\n",
+       "      <th>frame_num</th>\n",
+       "      <th>original_label</th>\n",
+       "      <th>scientific_name</th>\n",
+       "      <th>...</th>\n",
+       "      <th>superfamily</th>\n",
+       "      <th>family</th>\n",
+       "      <th>subfamily</th>\n",
+       "      <th>tribe</th>\n",
+       "      <th>genus</th>\n",
+       "      <th>species</th>\n",
+       "      <th>subspecies</th>\n",
+       "      <th>variety</th>\n",
+       "      <th>multi_species</th>\n",
+       "      <th>num_species</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>1369</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>...</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>iguanidae</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>False</td>\n",
+       "      <td>1.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>16627</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : chile_filip...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : chile_filip...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>petrel</td>\n",
+       "      <td>procellariidae</td>\n",
+       "      <td>...</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>procellariidae</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>False</td>\n",
+       "      <td>1.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>8315</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : chile_franc...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : chile_franc...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>rabbit</td>\n",
+       "      <td>leporidae</td>\n",
+       "      <td>...</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>leporidae</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>False</td>\n",
+       "      <td>1.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>7461</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : chile_franc...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : chile_franc...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>rabbit</td>\n",
+       "      <td>leporidae</td>\n",
+       "      <td>...</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>leporidae</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>False</td>\n",
+       "      <td>1.0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>4 rows × 34 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                           dataset_name  \\\n",
+       "1369   Island Conservation Camera Traps   \n",
+       "16627  Island Conservation Camera Traps   \n",
+       "8315   Island Conservation Camera Traps   \n",
+       "7461   Island Conservation Camera Traps   \n",
+       "\n",
+       "                                                 url_gcp  \\\n",
+       "1369   https://storage.googleapis.com/public-datasets...   \n",
+       "16627  https://storage.googleapis.com/public-datasets...   \n",
+       "8315   https://storage.googleapis.com/public-datasets...   \n",
+       "7461   https://storage.googleapis.com/public-datasets...   \n",
+       "\n",
+       "                                                 url_aws  \\\n",
+       "1369   http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "16627  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "8315   http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "7461   http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "\n",
+       "                                               url_azure  \\\n",
+       "1369   https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "16627  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "8315   https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "7461   https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "\n",
+       "                                                image_id  \\\n",
+       "1369   Island Conservation Camera Traps : dominicanre...   \n",
+       "16627  Island Conservation Camera Traps : chile_filip...   \n",
+       "8315   Island Conservation Camera Traps : chile_franc...   \n",
+       "7461   Island Conservation Camera Traps : chile_franc...   \n",
+       "\n",
+       "                                      sequence_id  \\\n",
+       "1369   Island Conservation Camera Traps : unknown   \n",
+       "16627  Island Conservation Camera Traps : unknown   \n",
+       "8315   Island Conservation Camera Traps : unknown   \n",
+       "7461   Island Conservation Camera Traps : unknown   \n",
+       "\n",
+       "                                             location_id  frame_num  \\\n",
+       "1369   Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "16627  Island Conservation Camera Traps : chile_filip...         -1   \n",
+       "8315   Island Conservation Camera Traps : chile_franc...         -1   \n",
+       "7461   Island Conservation Camera Traps : chile_franc...         -1   \n",
+       "\n",
+       "      original_label scientific_name  ... superfamily          family  \\\n",
+       "1369          iguana          iguana  ...         NaN       iguanidae   \n",
+       "16627         petrel  procellariidae  ...         NaN  procellariidae   \n",
+       "8315          rabbit       leporidae  ...         NaN       leporidae   \n",
+       "7461          rabbit       leporidae  ...         NaN       leporidae   \n",
+       "\n",
+       "      subfamily tribe   genus species subspecies variety multi_species  \\\n",
+       "1369        NaN   NaN  iguana     NaN        NaN     NaN         False   \n",
+       "16627       NaN   NaN     NaN     NaN        NaN     NaN         False   \n",
+       "8315        NaN   NaN     NaN     NaN        NaN     NaN         False   \n",
+       "7461        NaN   NaN     NaN     NaN        NaN     NaN         False   \n",
+       "\n",
+       "      num_species  \n",
+       "1369          1.0  \n",
+       "16627         1.0  \n",
+       "8315          1.0  \n",
+       "7461          1.0  \n",
+       "\n",
+       "[4 rows x 34 columns]"
+      ]
+     },
+     "execution_count": 17,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df.loc[df[\"species\"].isna()].sample(4)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Let's remove the multi-species images and set this up as a common name task."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 18,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "image_id           43577\n",
+       "scientific_name       41\n",
+       "common_name           43\n",
+       "species               22\n",
+       "dtype: int64"
+      ]
+     },
+     "execution_count": 18,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df_single = df.loc[~df[\"multi_species\"]].copy()\n",
+    "df_single[[\"image_id\",\"scientific_name\", \"common_name\", \"species\"]].nunique()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 19,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "common_name\n",
+       "rabbits and hares    7702\n",
+       "rat                  6844\n",
+       "petrel               6538\n",
+       "typical iguanas      6163\n",
+       "cat                  4663\n",
+       "goat                 2506\n",
+       "pig                  2283\n",
+       "shearwater           1699\n",
+       "petrel chick         1008\n",
+       "raven                 858\n",
+       "donkey                697\n",
+       "rooster               533\n",
+       "monitor lizard        354\n",
+       "megapode              274\n",
+       "white-winged dove     211\n",
+       "Name: count, dtype: int64"
+      ]
+     },
+     "execution_count": 19,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df_single.common_name.value_counts()[:15]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 20,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "common_name\n",
+       "crab                16\n",
+       "thrush              15\n",
+       "seal                15\n",
+       "storm petrel        12\n",
+       "nicobar pigeon      10\n",
+       "great blue heron     9\n",
+       "american kestrel     9\n",
+       "burrowing owl        6\n",
+       "sea turtle           6\n",
+       "green heron          6\n",
+       "spider               4\n",
+       "kestrel              4\n",
+       "mockingbird          3\n",
+       "barred owl           1\n",
+       "owl                  1\n",
+       "Name: count, dtype: int64"
+      ]
+     },
+     "execution_count": 20,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df_single.common_name.value_counts()[28:]"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "The other owl is \"short-eared owl\", if we combine them all to \"owl\" then we could keep them all. short-eared owls do seem to be well-represented though (see below). They are least concern but decreasing [IUCN](https://www.iucnredlist.org/species/22689531/202226582). Same for burrowing owls [IUCN](https://www.iucnredlist.org/species/22689353/93227732). Barred owls are both least concern and increasing [IUCN](https://www.iucnredlist.org/species/22689094/264599155), so combining may not be an issue (at least from that perspective)."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 21,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "193"
+      ]
+     },
+     "execution_count": 21,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df_single.loc[df_single[\"common_name\"] == \"short-eared owl\"].shape[0]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "location_id\n",
+       "Island Conservation Camera Traps : chile_filipiananbek           3647\n",
+       "Island Conservation Camera Traps : chile_frances01               2906\n",
+       "Island Conservation Camera Traps : chile_frances02               2800\n",
+       "Island Conservation Camera Traps : chile_piedra01                2788\n",
+       "Island Conservation Camera Traps : chile_filipianalamatris       2704\n",
+       "                                                                 ... \n",
+       "Island Conservation Camera Traps : ecuador1_cam1613                 5\n",
+       "Island Conservation Camera Traps : ecuador2_ic1602                  5\n",
+       "Island Conservation Camera Traps : dominicanrepublic_camara23       4\n",
+       "Island Conservation Camera Traps : palau_cam16a                     4\n",
+       "Island Conservation Camera Traps : ecuador2_ic1603                  4\n",
+       "Name: count, Length: 118, dtype: int64"
+      ]
+     },
+     "execution_count": 22,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df_single.location_id.value_counts()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 23,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "array(['Island Conservation Camera Traps : dominicanrepublic_camara02',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara04',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara06',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara101',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara103',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara107',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara109',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara111',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara114',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara13',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara17',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara20',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara30',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara10',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara106',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara112',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara11',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara116',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara12',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara19',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara113',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara15',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara08',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara108',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara118',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara18',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara14',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara33',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara05',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara07',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara09',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara105',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara32',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara01',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara03',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara117',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara24',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara27',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara16',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara102',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara104',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara29',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara115',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara22',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara28',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara119',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara120',\n",
+       "       'Island Conservation Camera Traps : dominicanrepublic_camara23',\n",
+       "       'Island Conservation Camera Traps : ecuador1_ic1619',\n",
+       "       'Island Conservation Camera Traps : ecuador1_ic1616',\n",
+       "       'Island Conservation Camera Traps : ecuador1_cam1612',\n",
+       "       'Island Conservation Camera Traps : ecuador1_cam1613',\n",
+       "       'Island Conservation Camera Traps : ecuador1_ic1611',\n",
+       "       'Island Conservation Camera Traps : chile_filipiananbek',\n",
+       "       'Island Conservation Camera Traps : chile_frances01',\n",
+       "       'Island Conservation Camera Traps : chile_frances02',\n",
+       "       'Island Conservation Camera Traps : chile_filipianalamatris',\n",
+       "       'Island Conservation Camera Traps : chile_piedra01',\n",
+       "       'Island Conservation Camera Traps : chile_vaqueria',\n",
+       "       'Island Conservation Camera Traps : puertorico_19',\n",
+       "       'Island Conservation Camera Traps : puertorico_20',\n",
+       "       'Island Conservation Camera Traps : puertorico_23',\n",
+       "       'Island Conservation Camera Traps : puertorico_25',\n",
+       "       'Island Conservation Camera Traps : puertorico_29',\n",
+       "       'Island Conservation Camera Traps : puertorico_3a',\n",
+       "       'Island Conservation Camera Traps : puertorico_4b',\n",
+       "       'Island Conservation Camera Traps : puertorico_2a',\n",
+       "       'Island Conservation Camera Traps : puertorico_7a',\n",
+       "       'Island Conservation Camera Traps : puertorico_26',\n",
+       "       'Island Conservation Camera Traps : puertorico_1a',\n",
+       "       'Island Conservation Camera Traps : puertorico_8b',\n",
+       "       'Island Conservation Camera Traps : puertorico_17b',\n",
+       "       'Island Conservation Camera Traps : puertorico_9a',\n",
+       "       'Island Conservation Camera Traps : puertorico_5a',\n",
+       "       'Island Conservation Camera Traps : puertorico_6a',\n",
+       "       'Island Conservation Camera Traps : puertorico_18',\n",
+       "       'Island Conservation Camera Traps : palau_cam09a',\n",
+       "       'Island Conservation Camera Traps : palau_cam02a',\n",
+       "       'Island Conservation Camera Traps : palau_cam04a',\n",
+       "       'Island Conservation Camera Traps : palau_cam05a',\n",
+       "       'Island Conservation Camera Traps : palau_cam06a',\n",
+       "       'Island Conservation Camera Traps : palau_cam10a',\n",
+       "       'Island Conservation Camera Traps : palau_cam13a',\n",
+       "       'Island Conservation Camera Traps : palau_cam14a',\n",
+       "       'Island Conservation Camera Traps : palau_cam15a',\n",
+       "       'Island Conservation Camera Traps : palau_cam16a',\n",
+       "       'Island Conservation Camera Traps : palau_cam17a',\n",
+       "       'Island Conservation Camera Traps : palau_cam01a',\n",
+       "       'Island Conservation Camera Traps : palau_cam03a',\n",
+       "       'Island Conservation Camera Traps : palau_cam07a',\n",
+       "       'Island Conservation Camera Traps : palau_cam08a',\n",
+       "       'Island Conservation Camera Traps : ecuador2_ic1618',\n",
+       "       'Island Conservation Camera Traps : ecuador2_ic1602',\n",
+       "       'Island Conservation Camera Traps : ecuador2_ic1603',\n",
+       "       'Island Conservation Camera Traps : ecuador2_ic1604',\n",
+       "       'Island Conservation Camera Traps : ecuador2_ic1605',\n",
+       "       'Island Conservation Camera Traps : ecuador2_ic1607',\n",
+       "       'Island Conservation Camera Traps : ecuador2_ic1609',\n",
+       "       'Island Conservation Camera Traps : ecuador2_ic1614',\n",
+       "       'Island Conservation Camera Traps : ecuador2_ic1621',\n",
+       "       'Island Conservation Camera Traps : ecuador2_ic1621a',\n",
+       "       'Island Conservation Camera Traps : micronesia_cam04',\n",
+       "       'Island Conservation Camera Traps : micronesia_cam12',\n",
+       "       'Island Conservation Camera Traps : micronesia_cam18',\n",
+       "       'Island Conservation Camera Traps : micronesia_cam19',\n",
+       "       'Island Conservation Camera Traps : micronesia_cam16',\n",
+       "       'Island Conservation Camera Traps : micronesia_cam10',\n",
+       "       'Island Conservation Camera Traps : micronesia_cam05',\n",
+       "       'Island Conservation Camera Traps : micronesia_cam13',\n",
+       "       'Island Conservation Camera Traps : micronesia_cam17',\n",
+       "       'Island Conservation Camera Traps : micronesia_cam15',\n",
+       "       'Island Conservation Camera Traps : micronesia_cam02',\n",
+       "       'Island Conservation Camera Traps : micronesia_cam03',\n",
+       "       'Island Conservation Camera Traps : micronesia_cam11',\n",
+       "       'Island Conservation Camera Traps : micronesia_cam14',\n",
+       "       'Island Conservation Camera Traps : micronesia_cam08',\n",
+       "       'Island Conservation Camera Traps : micronesia_cam06',\n",
+       "       'Island Conservation Camera Traps : micronesia_cam09'],\n",
+       "      dtype=object)"
+      ]
+     },
+     "execution_count": 23,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df_single.location_id.unique()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "The 6 countries are Dominican Republic, Ecuador, Chile, Puerto Rico, Palau, and Micronesia. It seems there are two islands in Ecuador."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Looking at least-represented species:\n",
+    " - mockingbirds have a variety of statuses, depending on the species, but some island mocking birds are vulnerable, endangered, or critically endangered [IUCN](https://www.iucnredlist.org/search?query=mockingbird&searchType=species).\n",
+    " - Similarly for kestrels.\n",
+    " - green heron not listed by IUCN\n",
+    " - sea turtles all vulnerable, endangered, critically endangered, or data deficient [IUCN](https://www.iucnredlist.org/search?query=Sea%20Turtles&searchType=species)."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "The `common_name` most prevalent in this dataset is \"rabbits and hares\", but these are two different species within the same family. Most of the images are classified down to the family level, so let's see what classification to family would look like as this may present a better set of classes, though we do have only 25 distinct families, compared to 45 common names."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 25,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "family\n",
+       "procellariidae    9245\n",
+       "leporidae         7702\n",
+       "muridae           6844\n",
+       "iguanidae         6163\n",
+       "felidae           4663\n",
+       "bovidae           2635\n",
+       "suidae            2283\n",
+       "corvidae           858\n",
+       "equidae            697\n",
+       "phasianidae        630\n",
+       "varanidae          354\n",
+       "megapodiidae       274\n",
+       "columbidae         250\n",
+       "strigidae          200\n",
+       "canidae            137\n",
+       "ardeidae           128\n",
+       "procyonidae        106\n",
+       "laridae            106\n",
+       "rallidae            51\n",
+       "calcinidae          39\n",
+       "turdidae            15\n",
+       "phocidae            15\n",
+       "falconidae          13\n",
+       "hydrobatidae        12\n",
+       "mimidae              3\n",
+       "Name: count, dtype: int64"
+      ]
+     },
+     "execution_count": 25,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df_single.family.value_counts()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "In this case, we would just lose mimidae (includes mocking birds) when creating a balanced set at 12 images per class. Perhaps we set up a balanced by-family task and a common name task that's imbalanced."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 24,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "<Axes: xlabel='Count', ylabel='family'>"
+      ]
+     },
+     "execution_count": 24,
+     "metadata": {},
+     "output_type": "execute_result"
+    },
+    {
+     "data": {
+      "image/png": "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",
+      "text/plain": [
+       "<Figure size 640x480 with 1 Axes>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "sns.histplot(df_single, y = 'family')"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### Common Name version\n",
+    "Replace \"rabbits and hares\" with \"rabbit or hare\" and \"typical iguanas\" with \"iguana\", then create the imbalanced and lower-bounded subsets.\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 27,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "df_common = df_single.copy()\n",
+    "df_common.loc[df_common[\"common_name\"] == \"rabbits and hares\", \"common_name\"] = \"rabbit or hare\"\n",
+    "df_common.loc[df_common[\"common_name\"] == \"typical iguanas\", \"common_name\"] = \"iguana\""
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### Filter to Family classifications"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 28,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "<class 'pandas.core.frame.DataFrame'>\n",
+      "Index: 43423 entries, 0 to 44437\n",
+      "Data columns (total 10 columns):\n",
+      " #   Column           Non-Null Count  Dtype \n",
+      "---  ------           --------------  ----- \n",
+      " 0   original_label   43423 non-null  object\n",
+      " 1   scientific_name  43423 non-null  object\n",
+      " 2   common_name      43423 non-null  object\n",
+      " 3   kingdom          43423 non-null  object\n",
+      " 4   phylum           43423 non-null  object\n",
+      " 5   class            43423 non-null  object\n",
+      " 6   order            43423 non-null  object\n",
+      " 7   family           43423 non-null  object\n",
+      " 8   genus            27449 non-null  object\n",
+      " 9   species          12728 non-null  object\n",
+      "dtypes: object(10)\n",
+      "memory usage: 3.6+ MB\n"
+     ]
+    }
+   ],
+   "source": [
+    "df_filter = df_single.loc[df_single[\"family\"].notna()].copy()\n",
+    "df_filter[taxa_cols].info(show_counts=True)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### Remove extra columns\n",
+    "\n",
+    "Only need `taxa_cols` (Linnean taxonomy + `original_label`, `scientific_name`, and `common_name`) and `id_cols`."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 29,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "id_cols = ['dataset_name',\n",
+    "           'url_gcp',\n",
+    "           'url_aws',\n",
+    "           'url_azure',\n",
+    "           'image_id',\n",
+    "           'sequence_id',\n",
+    "           'location_id',\n",
+    "           'frame_num']\n",
+    "\n",
+    "cols_to_keep = [col for col in list(df.columns) if (col in id_cols or col in taxa_cols)]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 30,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "['dataset_name',\n",
+       " 'url_gcp',\n",
+       " 'url_aws',\n",
+       " 'url_azure',\n",
+       " 'image_id',\n",
+       " 'sequence_id',\n",
+       " 'location_id',\n",
+       " 'frame_num',\n",
+       " 'original_label',\n",
+       " 'scientific_name',\n",
+       " 'common_name',\n",
+       " 'kingdom',\n",
+       " 'phylum',\n",
+       " 'class',\n",
+       " 'order',\n",
+       " 'family',\n",
+       " 'genus',\n",
+       " 'species']"
+      ]
+     },
+     "execution_count": 30,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "cols_to_keep"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Let's add a number of images column (by `family` and by `common_name` for each subset)."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 31,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>dataset_name</th>\n",
+       "      <th>url_gcp</th>\n",
+       "      <th>url_aws</th>\n",
+       "      <th>url_azure</th>\n",
+       "      <th>image_id</th>\n",
+       "      <th>sequence_id</th>\n",
+       "      <th>location_id</th>\n",
+       "      <th>frame_num</th>\n",
+       "      <th>original_label</th>\n",
+       "      <th>scientific_name</th>\n",
+       "      <th>...</th>\n",
+       "      <th>family</th>\n",
+       "      <th>subfamily</th>\n",
+       "      <th>tribe</th>\n",
+       "      <th>genus</th>\n",
+       "      <th>species</th>\n",
+       "      <th>subspecies</th>\n",
+       "      <th>variety</th>\n",
+       "      <th>multi_species</th>\n",
+       "      <th>num_species</th>\n",
+       "      <th>num_fam_images</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>cow</td>\n",
+       "      <td>bos taurus</td>\n",
+       "      <td>...</td>\n",
+       "      <td>bovidae</td>\n",
+       "      <td>bovinae</td>\n",
+       "      <td>bovini</td>\n",
+       "      <td>bos</td>\n",
+       "      <td>bos taurus</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>False</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>2635.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>donkey</td>\n",
+       "      <td>equus asinus</td>\n",
+       "      <td>...</td>\n",
+       "      <td>equidae</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>equus</td>\n",
+       "      <td>equus asinus</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>False</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>697.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>donkey</td>\n",
+       "      <td>equus asinus</td>\n",
+       "      <td>...</td>\n",
+       "      <td>equidae</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>equus</td>\n",
+       "      <td>equus asinus</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>False</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>697.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>...</td>\n",
+       "      <td>iguanidae</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>False</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>6163.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>...</td>\n",
+       "      <td>iguanidae</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>False</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>6163.0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>5 rows × 35 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                       dataset_name  \\\n",
+       "0  Island Conservation Camera Traps   \n",
+       "1  Island Conservation Camera Traps   \n",
+       "2  Island Conservation Camera Traps   \n",
+       "3  Island Conservation Camera Traps   \n",
+       "4  Island Conservation Camera Traps   \n",
+       "\n",
+       "                                             url_gcp  \\\n",
+       "0  https://storage.googleapis.com/public-datasets...   \n",
+       "1  https://storage.googleapis.com/public-datasets...   \n",
+       "2  https://storage.googleapis.com/public-datasets...   \n",
+       "3  https://storage.googleapis.com/public-datasets...   \n",
+       "4  https://storage.googleapis.com/public-datasets...   \n",
+       "\n",
+       "                                             url_aws  \\\n",
+       "0  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "1  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "2  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "3  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "4  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "\n",
+       "                                           url_azure  \\\n",
+       "0  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "1  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "2  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "3  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "4  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "\n",
+       "                                            image_id  \\\n",
+       "0  Island Conservation Camera Traps : dominicanre...   \n",
+       "1  Island Conservation Camera Traps : dominicanre...   \n",
+       "2  Island Conservation Camera Traps : dominicanre...   \n",
+       "3  Island Conservation Camera Traps : dominicanre...   \n",
+       "4  Island Conservation Camera Traps : dominicanre...   \n",
+       "\n",
+       "                                  sequence_id  \\\n",
+       "0  Island Conservation Camera Traps : unknown   \n",
+       "1  Island Conservation Camera Traps : unknown   \n",
+       "2  Island Conservation Camera Traps : unknown   \n",
+       "3  Island Conservation Camera Traps : unknown   \n",
+       "4  Island Conservation Camera Traps : unknown   \n",
+       "\n",
+       "                                         location_id  frame_num  \\\n",
+       "0  Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "1  Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "2  Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "3  Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "4  Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "\n",
+       "  original_label scientific_name  ...     family subfamily   tribe   genus  \\\n",
+       "0            cow      bos taurus  ...    bovidae   bovinae  bovini     bos   \n",
+       "1         donkey    equus asinus  ...    equidae       NaN     NaN   equus   \n",
+       "2         donkey    equus asinus  ...    equidae       NaN     NaN   equus   \n",
+       "3         iguana          iguana  ...  iguanidae       NaN     NaN  iguana   \n",
+       "4         iguana          iguana  ...  iguanidae       NaN     NaN  iguana   \n",
+       "\n",
+       "        species subspecies variety multi_species num_species num_fam_images  \n",
+       "0    bos taurus        NaN     NaN         False         1.0         2635.0  \n",
+       "1  equus asinus        NaN     NaN         False         1.0          697.0  \n",
+       "2  equus asinus        NaN     NaN         False         1.0          697.0  \n",
+       "3           NaN        NaN     NaN         False         1.0         6163.0  \n",
+       "4           NaN        NaN     NaN         False         1.0         6163.0  \n",
+       "\n",
+       "[5 rows x 35 columns]"
+      ]
+     },
+     "execution_count": 31,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "for sci_name in list(df_filter[\"family\"].unique()):\n",
+    "    df_filter.loc[df_filter[\"family\"] == sci_name, \"num_fam_images\"] = df_filter.loc[df_filter[\"family\"] == sci_name].shape[0]\n",
+    "\n",
+    "df_filter.head()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 32,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>dataset_name</th>\n",
+       "      <th>url_gcp</th>\n",
+       "      <th>url_aws</th>\n",
+       "      <th>url_azure</th>\n",
+       "      <th>image_id</th>\n",
+       "      <th>sequence_id</th>\n",
+       "      <th>location_id</th>\n",
+       "      <th>frame_num</th>\n",
+       "      <th>original_label</th>\n",
+       "      <th>scientific_name</th>\n",
+       "      <th>...</th>\n",
+       "      <th>family</th>\n",
+       "      <th>subfamily</th>\n",
+       "      <th>tribe</th>\n",
+       "      <th>genus</th>\n",
+       "      <th>species</th>\n",
+       "      <th>subspecies</th>\n",
+       "      <th>variety</th>\n",
+       "      <th>multi_species</th>\n",
+       "      <th>num_species</th>\n",
+       "      <th>num_cn_images</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>cow</td>\n",
+       "      <td>bos taurus</td>\n",
+       "      <td>...</td>\n",
+       "      <td>bovidae</td>\n",
+       "      <td>bovinae</td>\n",
+       "      <td>bovini</td>\n",
+       "      <td>bos</td>\n",
+       "      <td>bos taurus</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>False</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>129.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>donkey</td>\n",
+       "      <td>equus asinus</td>\n",
+       "      <td>...</td>\n",
+       "      <td>equidae</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>equus</td>\n",
+       "      <td>equus asinus</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>False</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>697.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>donkey</td>\n",
+       "      <td>equus asinus</td>\n",
+       "      <td>...</td>\n",
+       "      <td>equidae</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>equus</td>\n",
+       "      <td>equus asinus</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>False</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>697.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>...</td>\n",
+       "      <td>iguanidae</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>False</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>6163.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>...</td>\n",
+       "      <td>iguanidae</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>False</td>\n",
+       "      <td>1.0</td>\n",
+       "      <td>6163.0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "<p>5 rows × 35 columns</p>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                       dataset_name  \\\n",
+       "0  Island Conservation Camera Traps   \n",
+       "1  Island Conservation Camera Traps   \n",
+       "2  Island Conservation Camera Traps   \n",
+       "3  Island Conservation Camera Traps   \n",
+       "4  Island Conservation Camera Traps   \n",
+       "\n",
+       "                                             url_gcp  \\\n",
+       "0  https://storage.googleapis.com/public-datasets...   \n",
+       "1  https://storage.googleapis.com/public-datasets...   \n",
+       "2  https://storage.googleapis.com/public-datasets...   \n",
+       "3  https://storage.googleapis.com/public-datasets...   \n",
+       "4  https://storage.googleapis.com/public-datasets...   \n",
+       "\n",
+       "                                             url_aws  \\\n",
+       "0  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "1  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "2  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "3  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "4  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "\n",
+       "                                           url_azure  \\\n",
+       "0  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "1  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "2  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "3  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "4  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "\n",
+       "                                            image_id  \\\n",
+       "0  Island Conservation Camera Traps : dominicanre...   \n",
+       "1  Island Conservation Camera Traps : dominicanre...   \n",
+       "2  Island Conservation Camera Traps : dominicanre...   \n",
+       "3  Island Conservation Camera Traps : dominicanre...   \n",
+       "4  Island Conservation Camera Traps : dominicanre...   \n",
+       "\n",
+       "                                  sequence_id  \\\n",
+       "0  Island Conservation Camera Traps : unknown   \n",
+       "1  Island Conservation Camera Traps : unknown   \n",
+       "2  Island Conservation Camera Traps : unknown   \n",
+       "3  Island Conservation Camera Traps : unknown   \n",
+       "4  Island Conservation Camera Traps : unknown   \n",
+       "\n",
+       "                                         location_id  frame_num  \\\n",
+       "0  Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "1  Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "2  Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "3  Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "4  Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "\n",
+       "  original_label scientific_name  ...     family subfamily   tribe   genus  \\\n",
+       "0            cow      bos taurus  ...    bovidae   bovinae  bovini     bos   \n",
+       "1         donkey    equus asinus  ...    equidae       NaN     NaN   equus   \n",
+       "2         donkey    equus asinus  ...    equidae       NaN     NaN   equus   \n",
+       "3         iguana          iguana  ...  iguanidae       NaN     NaN  iguana   \n",
+       "4         iguana          iguana  ...  iguanidae       NaN     NaN  iguana   \n",
+       "\n",
+       "        species subspecies variety multi_species num_species num_cn_images  \n",
+       "0    bos taurus        NaN     NaN         False         1.0         129.0  \n",
+       "1  equus asinus        NaN     NaN         False         1.0         697.0  \n",
+       "2  equus asinus        NaN     NaN         False         1.0         697.0  \n",
+       "3           NaN        NaN     NaN         False         1.0        6163.0  \n",
+       "4           NaN        NaN     NaN         False         1.0        6163.0  \n",
+       "\n",
+       "[5 rows x 35 columns]"
+      ]
+     },
+     "execution_count": 32,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "for common_name in list(df_common[\"common_name\"].unique()):\n",
+    "    df_common.loc[df_common[\"common_name\"] == common_name, \"num_cn_images\"] = df_common.loc[df_common[\"common_name\"] == common_name].shape[0]\n",
+    "\n",
+    "df_common.head()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 33,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "cols_to_keep_common = [col for col in cols_to_keep]\n",
+    "cols_to_keep_common.append(\"num_cn_images\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 34,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "cols_to_keep.append(\"num_fam_images\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 35,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>dataset_name</th>\n",
+       "      <th>url_gcp</th>\n",
+       "      <th>url_aws</th>\n",
+       "      <th>url_azure</th>\n",
+       "      <th>image_id</th>\n",
+       "      <th>sequence_id</th>\n",
+       "      <th>location_id</th>\n",
+       "      <th>frame_num</th>\n",
+       "      <th>original_label</th>\n",
+       "      <th>scientific_name</th>\n",
+       "      <th>common_name</th>\n",
+       "      <th>kingdom</th>\n",
+       "      <th>phylum</th>\n",
+       "      <th>class</th>\n",
+       "      <th>order</th>\n",
+       "      <th>family</th>\n",
+       "      <th>genus</th>\n",
+       "      <th>species</th>\n",
+       "      <th>num_fam_images</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>cow</td>\n",
+       "      <td>bos taurus</td>\n",
+       "      <td>cattle</td>\n",
+       "      <td>animalia</td>\n",
+       "      <td>chordata</td>\n",
+       "      <td>mammalia</td>\n",
+       "      <td>artiodactyla</td>\n",
+       "      <td>bovidae</td>\n",
+       "      <td>bos</td>\n",
+       "      <td>bos taurus</td>\n",
+       "      <td>2635.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>donkey</td>\n",
+       "      <td>equus asinus</td>\n",
+       "      <td>donkey</td>\n",
+       "      <td>animalia</td>\n",
+       "      <td>chordata</td>\n",
+       "      <td>mammalia</td>\n",
+       "      <td>perissodactyla</td>\n",
+       "      <td>equidae</td>\n",
+       "      <td>equus</td>\n",
+       "      <td>equus asinus</td>\n",
+       "      <td>697.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>donkey</td>\n",
+       "      <td>equus asinus</td>\n",
+       "      <td>donkey</td>\n",
+       "      <td>animalia</td>\n",
+       "      <td>chordata</td>\n",
+       "      <td>mammalia</td>\n",
+       "      <td>perissodactyla</td>\n",
+       "      <td>equidae</td>\n",
+       "      <td>equus</td>\n",
+       "      <td>equus asinus</td>\n",
+       "      <td>697.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>typical iguanas</td>\n",
+       "      <td>animalia</td>\n",
+       "      <td>chordata</td>\n",
+       "      <td>reptilia</td>\n",
+       "      <td>squamata</td>\n",
+       "      <td>iguanidae</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>6163.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>typical iguanas</td>\n",
+       "      <td>animalia</td>\n",
+       "      <td>chordata</td>\n",
+       "      <td>reptilia</td>\n",
+       "      <td>squamata</td>\n",
+       "      <td>iguanidae</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>6163.0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                       dataset_name  \\\n",
+       "0  Island Conservation Camera Traps   \n",
+       "1  Island Conservation Camera Traps   \n",
+       "2  Island Conservation Camera Traps   \n",
+       "3  Island Conservation Camera Traps   \n",
+       "4  Island Conservation Camera Traps   \n",
+       "\n",
+       "                                             url_gcp  \\\n",
+       "0  https://storage.googleapis.com/public-datasets...   \n",
+       "1  https://storage.googleapis.com/public-datasets...   \n",
+       "2  https://storage.googleapis.com/public-datasets...   \n",
+       "3  https://storage.googleapis.com/public-datasets...   \n",
+       "4  https://storage.googleapis.com/public-datasets...   \n",
+       "\n",
+       "                                             url_aws  \\\n",
+       "0  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "1  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "2  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "3  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "4  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "\n",
+       "                                           url_azure  \\\n",
+       "0  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "1  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "2  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "3  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "4  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "\n",
+       "                                            image_id  \\\n",
+       "0  Island Conservation Camera Traps : dominicanre...   \n",
+       "1  Island Conservation Camera Traps : dominicanre...   \n",
+       "2  Island Conservation Camera Traps : dominicanre...   \n",
+       "3  Island Conservation Camera Traps : dominicanre...   \n",
+       "4  Island Conservation Camera Traps : dominicanre...   \n",
+       "\n",
+       "                                  sequence_id  \\\n",
+       "0  Island Conservation Camera Traps : unknown   \n",
+       "1  Island Conservation Camera Traps : unknown   \n",
+       "2  Island Conservation Camera Traps : unknown   \n",
+       "3  Island Conservation Camera Traps : unknown   \n",
+       "4  Island Conservation Camera Traps : unknown   \n",
+       "\n",
+       "                                         location_id  frame_num  \\\n",
+       "0  Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "1  Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "2  Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "3  Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "4  Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "\n",
+       "  original_label scientific_name      common_name   kingdom    phylum  \\\n",
+       "0            cow      bos taurus           cattle  animalia  chordata   \n",
+       "1         donkey    equus asinus           donkey  animalia  chordata   \n",
+       "2         donkey    equus asinus           donkey  animalia  chordata   \n",
+       "3         iguana          iguana  typical iguanas  animalia  chordata   \n",
+       "4         iguana          iguana  typical iguanas  animalia  chordata   \n",
+       "\n",
+       "      class           order     family   genus       species  num_fam_images  \n",
+       "0  mammalia    artiodactyla    bovidae     bos    bos taurus          2635.0  \n",
+       "1  mammalia  perissodactyla    equidae   equus  equus asinus           697.0  \n",
+       "2  mammalia  perissodactyla    equidae   equus  equus asinus           697.0  \n",
+       "3  reptilia        squamata  iguanidae  iguana           NaN          6163.0  \n",
+       "4  reptilia        squamata  iguanidae  iguana           NaN          6163.0  "
+      ]
+     },
+     "execution_count": 35,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df_reduced = df_filter[cols_to_keep].copy()\n",
+    "df_reduced.head()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 36,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>dataset_name</th>\n",
+       "      <th>url_gcp</th>\n",
+       "      <th>url_aws</th>\n",
+       "      <th>url_azure</th>\n",
+       "      <th>image_id</th>\n",
+       "      <th>sequence_id</th>\n",
+       "      <th>location_id</th>\n",
+       "      <th>frame_num</th>\n",
+       "      <th>original_label</th>\n",
+       "      <th>scientific_name</th>\n",
+       "      <th>common_name</th>\n",
+       "      <th>kingdom</th>\n",
+       "      <th>phylum</th>\n",
+       "      <th>class</th>\n",
+       "      <th>order</th>\n",
+       "      <th>family</th>\n",
+       "      <th>genus</th>\n",
+       "      <th>species</th>\n",
+       "      <th>num_cn_images</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>cow</td>\n",
+       "      <td>bos taurus</td>\n",
+       "      <td>cattle</td>\n",
+       "      <td>animalia</td>\n",
+       "      <td>chordata</td>\n",
+       "      <td>mammalia</td>\n",
+       "      <td>artiodactyla</td>\n",
+       "      <td>bovidae</td>\n",
+       "      <td>bos</td>\n",
+       "      <td>bos taurus</td>\n",
+       "      <td>129.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>donkey</td>\n",
+       "      <td>equus asinus</td>\n",
+       "      <td>donkey</td>\n",
+       "      <td>animalia</td>\n",
+       "      <td>chordata</td>\n",
+       "      <td>mammalia</td>\n",
+       "      <td>perissodactyla</td>\n",
+       "      <td>equidae</td>\n",
+       "      <td>equus</td>\n",
+       "      <td>equus asinus</td>\n",
+       "      <td>697.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>donkey</td>\n",
+       "      <td>equus asinus</td>\n",
+       "      <td>donkey</td>\n",
+       "      <td>animalia</td>\n",
+       "      <td>chordata</td>\n",
+       "      <td>mammalia</td>\n",
+       "      <td>perissodactyla</td>\n",
+       "      <td>equidae</td>\n",
+       "      <td>equus</td>\n",
+       "      <td>equus asinus</td>\n",
+       "      <td>697.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>animalia</td>\n",
+       "      <td>chordata</td>\n",
+       "      <td>reptilia</td>\n",
+       "      <td>squamata</td>\n",
+       "      <td>iguanidae</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>6163.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>Island Conservation Camera Traps</td>\n",
+       "      <td>https://storage.googleapis.com/public-datasets...</td>\n",
+       "      <td>http://us-west-2.opendata.source.coop.s3.amazo...</td>\n",
+       "      <td>https://lilawildlife.blob.core.windows.net/lil...</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>Island Conservation Camera Traps : unknown</td>\n",
+       "      <td>Island Conservation Camera Traps : dominicanre...</td>\n",
+       "      <td>-1</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>animalia</td>\n",
+       "      <td>chordata</td>\n",
+       "      <td>reptilia</td>\n",
+       "      <td>squamata</td>\n",
+       "      <td>iguanidae</td>\n",
+       "      <td>iguana</td>\n",
+       "      <td>NaN</td>\n",
+       "      <td>6163.0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "                       dataset_name  \\\n",
+       "0  Island Conservation Camera Traps   \n",
+       "1  Island Conservation Camera Traps   \n",
+       "2  Island Conservation Camera Traps   \n",
+       "3  Island Conservation Camera Traps   \n",
+       "4  Island Conservation Camera Traps   \n",
+       "\n",
+       "                                             url_gcp  \\\n",
+       "0  https://storage.googleapis.com/public-datasets...   \n",
+       "1  https://storage.googleapis.com/public-datasets...   \n",
+       "2  https://storage.googleapis.com/public-datasets...   \n",
+       "3  https://storage.googleapis.com/public-datasets...   \n",
+       "4  https://storage.googleapis.com/public-datasets...   \n",
+       "\n",
+       "                                             url_aws  \\\n",
+       "0  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "1  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "2  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "3  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "4  http://us-west-2.opendata.source.coop.s3.amazo...   \n",
+       "\n",
+       "                                           url_azure  \\\n",
+       "0  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "1  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "2  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "3  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "4  https://lilawildlife.blob.core.windows.net/lil...   \n",
+       "\n",
+       "                                            image_id  \\\n",
+       "0  Island Conservation Camera Traps : dominicanre...   \n",
+       "1  Island Conservation Camera Traps : dominicanre...   \n",
+       "2  Island Conservation Camera Traps : dominicanre...   \n",
+       "3  Island Conservation Camera Traps : dominicanre...   \n",
+       "4  Island Conservation Camera Traps : dominicanre...   \n",
+       "\n",
+       "                                  sequence_id  \\\n",
+       "0  Island Conservation Camera Traps : unknown   \n",
+       "1  Island Conservation Camera Traps : unknown   \n",
+       "2  Island Conservation Camera Traps : unknown   \n",
+       "3  Island Conservation Camera Traps : unknown   \n",
+       "4  Island Conservation Camera Traps : unknown   \n",
+       "\n",
+       "                                         location_id  frame_num  \\\n",
+       "0  Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "1  Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "2  Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "3  Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "4  Island Conservation Camera Traps : dominicanre...         -1   \n",
+       "\n",
+       "  original_label scientific_name common_name   kingdom    phylum     class  \\\n",
+       "0            cow      bos taurus      cattle  animalia  chordata  mammalia   \n",
+       "1         donkey    equus asinus      donkey  animalia  chordata  mammalia   \n",
+       "2         donkey    equus asinus      donkey  animalia  chordata  mammalia   \n",
+       "3         iguana          iguana      iguana  animalia  chordata  reptilia   \n",
+       "4         iguana          iguana      iguana  animalia  chordata  reptilia   \n",
+       "\n",
+       "            order     family   genus       species  num_cn_images  \n",
+       "0    artiodactyla    bovidae     bos    bos taurus          129.0  \n",
+       "1  perissodactyla    equidae   equus  equus asinus          697.0  \n",
+       "2  perissodactyla    equidae   equus  equus asinus          697.0  \n",
+       "3        squamata  iguanidae  iguana           NaN         6163.0  \n",
+       "4        squamata  iguanidae  iguana           NaN         6163.0  "
+      ]
+     },
+     "execution_count": 36,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df_reduced_common = df_common[cols_to_keep_common].copy()\n",
+    "df_reduced_common.head()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### Reduce to no more than 10K images per species"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 37,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "common_name\n",
+       "rabbit or hare    7702\n",
+       "rat               6844\n",
+       "petrel            6538\n",
+       "iguana            6163\n",
+       "cat               4663\n",
+       "Name: count, dtype: int64"
+      ]
+     },
+     "execution_count": 37,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df_reduced_common[\"common_name\"].value_counts()[:5]"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 38,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "family\n",
+       "procellariidae    9245\n",
+       "leporidae         7702\n",
+       "muridae           6844\n",
+       "iguanidae         6163\n",
+       "felidae           4663\n",
+       "Name: count, dtype: int64"
+      ]
+     },
+     "execution_count": 38,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "df_reduced.family.value_counts()[:5]"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Both subsets are already under 10K."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 39,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "df_reduced_common.to_csv(\"../data/potential-test-sets/filtered/island-imbalanced_common.csv\", index = False)\n",
+    "df_reduced.to_csv(\"../data/potential-test-sets/filtered/island-imbalanced_family.csv\", index = False)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 40,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "df_reduced_common.loc[df_reduced_common[\"num_cn_images\"] >= 10].to_csv(\"../data/potential-test-sets/filtered/island-lower-bound_common.csv\", index = False)\n",
+    "df_reduced.loc[df_reduced[\"num_fam_images\"] >= 10].to_csv(\"../data/potential-test-sets/filtered/island-lower-bound_family.csv\", index = False)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 42,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "<Axes: xlabel='Count', ylabel='common_name'>"
+      ]
+     },
+     "execution_count": 42,
+     "metadata": {},
+     "output_type": "execute_result"
+    },
+    {
+     "data": {
+      "image/png": "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",
+      "text/plain": [
+       "<Figure size 640x480 with 1 Axes>"
+      ]
+     },
+     "metadata": {},
+     "output_type": "display_data"
+    }
+   ],
+   "source": [
+    "sns.histplot(df_reduced_common.loc[df_reduced_common[\"num_cn_images\"] >= 10], y = \"common_name\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### Randomly sample to balanced set (12 images per family)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 43,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "288"
+      ]
+     },
+     "execution_count": 43,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "balanced_set = []\n",
+    "for sci_name in list(df_reduced[\"family\"].unique()):\n",
+    "    temp = df_reduced.loc[df_reduced[\"family\"] == sci_name].copy()\n",
+    "    if temp.shape[0] < 12:\n",
+    "        continue\n",
+    "    sample_set = list(temp.sample(12, random_state = 614)[\"image_id\"])\n",
+    "    balanced_set = balanced_set + sample_set\n",
+    "\n",
+    "len(balanced_set)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Filter to just balanced set and drop the number of species column since it's been balanced to 12 each."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 45,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "df_balanced = df_reduced.loc[df_reduced[\"image_id\"].isin(balanced_set)].copy()\n",
+    "df_balanced.drop(columns = [\"num_fam_images\"], inplace = True)\n",
+    "df_balanced.to_csv(\"../data/potential-test-sets/filtered/island-balanced.csv\", index = False)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "jupytext": {
+   "formats": "ipynb,py:percent"
+  },
+  "kernelspec": {
+   "display_name": "data-dev",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.12.3"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}