diff --git "a/notebooks/ToL_EDA.ipynb" "b/notebooks/ToL_EDA.ipynb" new file mode 100644--- /dev/null +++ "b/notebooks/ToL_EDA.ipynb" @@ -0,0 +1,2713 @@ +{ + "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\")\n", + "sns.set(rc = {'figure.figsize': (10,10)})" + ] + }, + { + "cell_type": "code", + "execution_count": 2, + "metadata": {}, + "outputs": [ + { + "name": "stderr", + "output_type": "stream", + "text": [ + "/tmp/ipykernel_414099/3694103411.py:1: DtypeWarning: Columns (4,5,6) have mixed types. Specify dtype option on import or set low_memory=False.\n", + " df = pd.read_csv(\"../data/v1-dev-names.csv\")\n" + ] + } + ], + "source": [ + "df = pd.read_csv(\"../data/v1-dev-names.csv\")" + ] + }, + { + "cell_type": "code", + "execution_count": 3, + "metadata": {}, + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
treeoflife_ideol_content_ideol_page_idbioscan_partbioscan_filenameinat21_filenameinat21_cls_nameinat21_cls_numkingdomphylumclassorderfamilygenusspeciescommon
00824741f-cc1c-4881-b292-15fd3f7964cd29538374.065414274.0NaNNaNNaNNaNNaNNaNNaNNaNNaNNaNManfredaNaNtuberose
15ca08f6b-9396-4cb9-9283-8dee158aac1827793900.0888015.0NaNNaNNaNNaNNaNMetazoaArthropodaPancrustaceaLepidopteraOenosandridaeDiscophlebialipaugesDiscophlebia lipauges
2f8c0f271-d8e5-4299-92d3-920508f74bf029121641.05618956.0NaNNaNNaNNaNNaNArchaeplastidaTracheophytaNaNSapindalesRutaceaeMelicopedenhamiiMelicope denhamii
31f53e9d1-527f-42fd-b813-9f62fa2c237227596176.0607817.0NaNNaNNaNNaNNaNMetazoaArthropodaPancrustaceaTrichopteraLimnephilidaeLimnephiluslithusLimnephilus lithus
4a05bc2a8-5453-4683-903e-ed44f0fe724520300703.0267922.0NaNNaNNaNNaNNaNNaNNaNNaNNaNNaNNaNNaNAnatolian Black-eyed Blue
\n", + "
" + ], + "text/plain": [ + " treeoflife_id eol_content_id eol_page_id \\\n", + "0 0824741f-cc1c-4881-b292-15fd3f7964cd 29538374.0 65414274.0 \n", + "1 5ca08f6b-9396-4cb9-9283-8dee158aac18 27793900.0 888015.0 \n", + "2 f8c0f271-d8e5-4299-92d3-920508f74bf0 29121641.0 5618956.0 \n", + "3 1f53e9d1-527f-42fd-b813-9f62fa2c2372 27596176.0 607817.0 \n", + "4 a05bc2a8-5453-4683-903e-ed44f0fe7245 20300703.0 267922.0 \n", + "\n", + " bioscan_part bioscan_filename inat21_filename inat21_cls_name \\\n", + "0 NaN NaN NaN NaN \n", + "1 NaN NaN NaN NaN \n", + "2 NaN NaN NaN NaN \n", + "3 NaN NaN NaN NaN \n", + "4 NaN NaN NaN NaN \n", + "\n", + " inat21_cls_num kingdom phylum class order \\\n", + "0 NaN NaN NaN NaN NaN \n", + "1 NaN Metazoa Arthropoda Pancrustacea Lepidoptera \n", + "2 NaN Archaeplastida Tracheophyta NaN Sapindales \n", + "3 NaN Metazoa Arthropoda Pancrustacea Trichoptera \n", + "4 NaN NaN NaN NaN NaN \n", + "\n", + " family genus species common \n", + "0 NaN Manfreda NaN tuberose \n", + "1 Oenosandridae Discophlebia lipauges Discophlebia lipauges \n", + "2 Rutaceae Melicope denhamii Melicope denhamii \n", + "3 Limnephilidae Limnephilus lithus Limnephilus lithus \n", + "4 NaN NaN NaN Anatolian Black-eyed Blue " + ] + }, + "execution_count": 3, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "df.head()" + ] + }, + { + "cell_type": "code", + "execution_count": 4, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "RangeIndex: 10436521 entries, 0 to 10436520\n", + "Data columns (total 16 columns):\n", + " # Column Non-Null Count Dtype \n", + "--- ------ -------------- ----- \n", + " 0 treeoflife_id 10436521 non-null object \n", + " 1 eol_content_id 6621365 non-null float64\n", + " 2 eol_page_id 6621365 non-null float64\n", + " 3 bioscan_part 1128313 non-null float64\n", + " 4 bioscan_filename 1128313 non-null object \n", + " 5 inat21_filename 2686843 non-null object \n", + " 6 inat21_cls_name 2686843 non-null object \n", + " 7 inat21_cls_num 2686843 non-null float64\n", + " 8 kingdom 7734559 non-null object \n", + " 9 phylum 7732689 non-null object \n", + " 10 class 6657484 non-null object \n", + " 11 order 7690349 non-null object \n", + " 12 family 7706759 non-null object \n", + " 13 genus 8060829 non-null object \n", + " 14 species 7179863 non-null object \n", + " 15 common 10436521 non-null object \n", + "dtypes: float64(4), object(12)\n", + "memory usage: 1.2+ GB\n" + ] + } + ], + "source": [ + "df.info(show_counts = True)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "Do we truly have common name labeled for all images?\n", + "\n", + "Sometimes it is the scientific name." + ] + }, + { + "cell_type": "code", + "execution_count": 5, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "treeoflife_id 10436521\n", + "eol_content_id 6621365\n", + "eol_page_id 570515\n", + "bioscan_part 113\n", + "bioscan_filename 1128313\n", + "inat21_filename 2686843\n", + "inat21_cls_name 10000\n", + "inat21_cls_num 10000\n", + "kingdom 5\n", + "phylum 49\n", + "class 136\n", + "order 766\n", + "family 5665\n", + "genus 89914\n", + "species 189846\n", + "common 527316\n", + "dtype: int64" + ] + }, + "execution_count": 5, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "df.nunique()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "There are 570,515 unique EOL page IDs, suggesting 570,515 unique classes among the 6,621,365 images pulled from EOL. Presumably this would represent the number of species or other lowest rank taxa covered." + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "Notice that we have 5 unique kingdoms, when there are only 3..." + ] + }, + { + "cell_type": "code", + "execution_count": 6, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "kingdom\n", + "Animalia 2576406\n", + "Metazoa 2207522\n", + "Archaeplastida 1481804\n", + "Plantae 1148702\n", + "Fungi 320125\n", + "Name: count, dtype: int64" + ] + }, + "execution_count": 6, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "df['kingdom'].value_counts()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "`Metazoa` and `Animalia` overlap, as do `Archaeplastida` and `Plantae`. They are sometimes used interchangably, though the former of each is a newer (more refined?) designation. Later we'll see this distinction cuts is by our data sources." + ] + }, + { + "cell_type": "code", + "execution_count": 7, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "['kingdom', 'phylum', 'class', 'order', 'family', 'genus', 'species']" + ] + }, + "execution_count": 7, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "taxa = list(df.columns[8:15])\n", + "taxa" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "Check the number of images with all 7 taxonomic labels." + ] + }, + { + "cell_type": "code", + "execution_count": 8, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "Index: 6657484 entries, 1 to 10436518\n", + "Data columns (total 7 columns):\n", + " # Column Non-Null Count Dtype \n", + "--- ------ -------------- ----- \n", + " 0 kingdom 6657484 non-null object\n", + " 1 phylum 6657484 non-null object\n", + " 2 class 6657484 non-null object\n", + " 3 order 6615972 non-null object\n", + " 4 family 6630424 non-null object\n", + " 5 genus 5691848 non-null object\n", + " 6 species 5267629 non-null object\n", + "dtypes: object(7)\n", + "memory usage: 406.3+ MB\n" + ] + } + ], + "source": [ + "# Class has least non-null entries, so we'll start by filtering it out\n", + "df_all_taxa = df.loc[df['class'].notna()]\n", + "df_all_taxa[taxa].info(show_counts = True)" + ] + }, + { + "cell_type": "code", + "execution_count": 9, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "Index: 5267629 entries, 1 to 10436518\n", + "Data columns (total 7 columns):\n", + " # Column Non-Null Count Dtype \n", + "--- ------ -------------- ----- \n", + " 0 kingdom 5267629 non-null object\n", + " 1 phylum 5267629 non-null object\n", + " 2 class 5267629 non-null object\n", + " 3 order 5233623 non-null object\n", + " 4 family 5261785 non-null object\n", + " 5 genus 5267456 non-null object\n", + " 6 species 5267629 non-null object\n", + "dtypes: object(7)\n", + "memory usage: 321.5+ MB\n" + ] + } + ], + "source": [ + "# Now species has least non-null values\n", + "df_all_taxa = df_all_taxa.loc[df_all_taxa['species'].notna()]\n", + "df_all_taxa[taxa].info(show_counts = True)" + ] + }, + { + "cell_type": "code", + "execution_count": 10, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "Index: 5233623 entries, 1 to 10436518\n", + "Data columns (total 7 columns):\n", + " # Column Non-Null Count Dtype \n", + "--- ------ -------------- ----- \n", + " 0 kingdom 5233623 non-null object\n", + " 1 phylum 5233623 non-null object\n", + " 2 class 5233623 non-null object\n", + " 3 order 5233623 non-null object\n", + " 4 family 5228644 non-null object\n", + " 5 genus 5233450 non-null object\n", + " 6 species 5233623 non-null object\n", + "dtypes: object(7)\n", + "memory usage: 319.4+ MB\n" + ] + } + ], + "source": [ + "# Now order has least non-null values \n", + "df_all_taxa = df_all_taxa.loc[df_all_taxa['order'].notna()]\n", + "df_all_taxa[taxa].info(show_counts = True)" + ] + }, + { + "cell_type": "code", + "execution_count": 11, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "Index: 5228644 entries, 1 to 10436518\n", + "Data columns (total 7 columns):\n", + " # Column Non-Null Count Dtype \n", + "--- ------ -------------- ----- \n", + " 0 kingdom 5228644 non-null object\n", + " 1 phylum 5228644 non-null object\n", + " 2 class 5228644 non-null object\n", + " 3 order 5228644 non-null object\n", + " 4 family 5228644 non-null object\n", + " 5 genus 5228471 non-null object\n", + " 6 species 5228644 non-null object\n", + "dtypes: object(7)\n", + "memory usage: 319.1+ MB\n" + ] + } + ], + "source": [ + "# Now family has least non-null values\n", + "df_all_taxa = df_all_taxa.loc[df_all_taxa['family'].notna()]\n", + "df_all_taxa[taxa].info(show_counts = True)" + ] + }, + { + "cell_type": "code", + "execution_count": 12, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "Index: 5228471 entries, 1 to 10436518\n", + "Data columns (total 7 columns):\n", + " # Column Non-Null Count Dtype \n", + "--- ------ -------------- ----- \n", + " 0 kingdom 5228471 non-null object\n", + " 1 phylum 5228471 non-null object\n", + " 2 class 5228471 non-null object\n", + " 3 order 5228471 non-null object\n", + " 4 family 5228471 non-null object\n", + " 5 genus 5228471 non-null object\n", + " 6 species 5228471 non-null object\n", + "dtypes: object(7)\n", + "memory usage: 319.1+ MB\n" + ] + } + ], + "source": [ + "# Finally, genus is the only one with null values\n", + "df_all_taxa = df_all_taxa.loc[df_all_taxa['genus'].notna()]\n", + "df_all_taxa[taxa].info(show_counts = True)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "We have 5,228,471 images with full taxonomic labels.\n", + "\n", + "Notice that we had gaps in the taxonomic hierarchy (both higher and lower values). This can (and should) be fixed using our taxon matching to get the higher order ranks for every entry with at least genus-species designation, implying that we should have at least 7M entries with full taxonomic labels." + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "Let's add a column indicating the original data source so we can also get some stats by datasource." + ] + }, + { + "cell_type": "code", + "execution_count": 13, + "metadata": {}, + "outputs": [], + "source": [ + "# Add data_source column for easier slicing\n", + "df.loc[df['inat21_filename'].notna(), 'data_source'] = 'iNat21'\n", + "df.loc[df['bioscan_filename'].notna(), 'data_source'] = 'BIOSCAN'\n", + "df.loc[df['eol_content_id'].notna(), 'data_source'] = 'EOL'" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "First, check their unique class values (`common`)." + ] + }, + { + "cell_type": "code", + "execution_count": 14, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "518118" + ] + }, + "execution_count": 14, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "df.loc[df['data_source'] == 'EOL', 'common'].nunique()" + ] + }, + { + "cell_type": "code", + "execution_count": 15, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "9941" + ] + }, + "execution_count": 15, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "df.loc[df['data_source'] == 'iNat21', 'common'].nunique()" + ] + }, + { + "cell_type": "code", + "execution_count": 16, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "10635" + ] + }, + "execution_count": 16, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "df.loc[df['data_source'] == 'BIOSCAN', 'common'].nunique()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "Make `df_taxa` with just taxa columns (+ `common` & `data_source`) so it's smaller to process faster." + ] + }, + { + "cell_type": "code", + "execution_count": 17, + "metadata": {}, + "outputs": [], + "source": [ + "taxa_com = list(df.columns[8:16]) # taxa + common\n", + "taxa_com.insert(0, 'data_source')\n", + "df_taxa = df[taxa_com]" + ] + }, + { + "cell_type": "code", + "execution_count": 18, + "metadata": {}, + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
data_sourcekingdomphylumclassorderfamilygenusspeciescommon
0EOLNaNNaNNaNNaNNaNManfredaNaNtuberose
1EOLMetazoaArthropodaPancrustaceaLepidopteraOenosandridaeDiscophlebialipaugesDiscophlebia lipauges
2EOLArchaeplastidaTracheophytaNaNSapindalesRutaceaeMelicopedenhamiiMelicope denhamii
3EOLMetazoaArthropodaPancrustaceaTrichopteraLimnephilidaeLimnephiluslithusLimnephilus lithus
4EOLNaNNaNNaNNaNNaNNaNNaNAnatolian Black-eyed Blue
\n", + "
" + ], + "text/plain": [ + " data_source kingdom phylum class order \\\n", + "0 EOL NaN NaN NaN NaN \n", + "1 EOL Metazoa Arthropoda Pancrustacea Lepidoptera \n", + "2 EOL Archaeplastida Tracheophyta NaN Sapindales \n", + "3 EOL Metazoa Arthropoda Pancrustacea Trichoptera \n", + "4 EOL NaN NaN NaN NaN \n", + "\n", + " family genus species common \n", + "0 NaN Manfreda NaN tuberose \n", + "1 Oenosandridae Discophlebia lipauges Discophlebia lipauges \n", + "2 Rutaceae Melicope denhamii Melicope denhamii \n", + "3 Limnephilidae Limnephilus lithus Limnephilus lithus \n", + "4 NaN NaN NaN Anatolian Black-eyed Blue " + ] + }, + "execution_count": 18, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "df_taxa.head()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "Notice that the third entry is missing class. Do we have issues like this in BIOSCAN or iNat21? We shouldn't." + ] + }, + { + "cell_type": "code", + "execution_count": 19, + "metadata": {}, + "outputs": [], + "source": [ + "inat21_df = df_taxa.loc[df_taxa.data_source == 'iNat21']\n", + "bioscan_df = df_taxa.loc[df_taxa.data_source == 'BIOSCAN']\n", + "eol_df = df_taxa.loc[df_taxa.data_source == 'EOL']" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "### iNat21" + ] + }, + { + "cell_type": "code", + "execution_count": 20, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "Index: 2686843 entries, 4523720 to 10327558\n", + "Data columns (total 9 columns):\n", + " # Column Non-Null Count Dtype \n", + "--- ------ -------------- ----- \n", + " 0 data_source 2686843 non-null object\n", + " 1 kingdom 2686843 non-null object\n", + " 2 phylum 2686843 non-null object\n", + " 3 class 2686843 non-null object\n", + " 4 order 2686843 non-null object\n", + " 5 family 2686843 non-null object\n", + " 6 genus 2686843 non-null object\n", + " 7 species 2686843 non-null object\n", + " 8 common 2686843 non-null object\n", + "dtypes: object(9)\n", + "memory usage: 205.0+ MB\n" + ] + } + ], + "source": [ + "inat21_df.info(show_counts = True)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "iNat21 isn't missing anything, as expected, and we have 2,686,843 images.\n", + "\n", + "Quick view of diversity in iNat21." + ] + }, + { + "cell_type": "code", + "execution_count": 21, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "data_source 1\n", + "kingdom 3\n", + "phylum 13\n", + "class 51\n", + "order 273\n", + "family 1103\n", + "genus 4884\n", + "species 6485\n", + "common 9941\n", + "dtype: int64" + ] + }, + "execution_count": 21, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "inat21_df.nunique()" + ] + }, + { + "cell_type": "code", + "execution_count": 22, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "kingdom\n", + "Animalia 1448093\n", + "Plantae 1148702\n", + "Fungi 90048\n", + "Name: count, dtype: int64" + ] + }, + "execution_count": 22, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "inat21_df['kingdom'].value_counts()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "iNat21 uses `Animalia` and `Plantae`." + ] + }, + { + "cell_type": "code", + "execution_count": 23, + "metadata": {}, + "outputs": [ + { + "name": "stderr", + "output_type": "stream", + "text": [ + "/tmp/ipykernel_414099/2358816078.py:2: SettingWithCopyWarning: \n", + "A value is trying to be set on a copy of a slice from a DataFrame.\n", + "Try using .loc[row_indexer,col_indexer] = value instead\n", + "\n", + "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n", + " inat21_df['duplicate'] = inat21_df.duplicated(subset = taxa, keep = 'first')\n" + ] + } + ], + "source": [ + "#number of unique 7-tuples in iNat21\n", + "inat21_df['duplicate'] = inat21_df.duplicated(subset = taxa, keep = 'first')\n", + "inat21_df_unique_taxa = inat21_df.loc[~inat21_df['duplicate']]" + ] + }, + { + "cell_type": "code", + "execution_count": 24, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "Index: 10000 entries, 4523720 to 10327259\n", + "Data columns (total 10 columns):\n", + " # Column Non-Null Count Dtype \n", + "--- ------ -------------- ----- \n", + " 0 data_source 10000 non-null object\n", + " 1 kingdom 10000 non-null object\n", + " 2 phylum 10000 non-null object\n", + " 3 class 10000 non-null object\n", + " 4 order 10000 non-null object\n", + " 5 family 10000 non-null object\n", + " 6 genus 10000 non-null object\n", + " 7 species 10000 non-null object\n", + " 8 common 10000 non-null object\n", + " 9 duplicate 10000 non-null bool \n", + "dtypes: bool(1), object(9)\n", + "memory usage: 791.0+ KB\n" + ] + } + ], + "source": [ + "inat21_df_unique_taxa.info(show_counts = True)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "There's the 10K unique `species` that we expect. Let's check the same information across BIOSCAN and EOL." + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "### BIOSCAN" + ] + }, + { + "cell_type": "code", + "execution_count": 25, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "Index: 1128313 entries, 4319650 to 8403834\n", + "Data columns (total 9 columns):\n", + " # Column Non-Null Count Dtype \n", + "--- ------ -------------- ----- \n", + " 0 data_source 1128313 non-null object\n", + " 1 kingdom 1128313 non-null object\n", + " 2 phylum 1128313 non-null object\n", + " 3 class 1128313 non-null object\n", + " 4 order 1128313 non-null object\n", + " 5 family 1112922 non-null object\n", + " 6 genus 254149 non-null object\n", + " 7 species 84450 non-null object\n", + " 8 common 1128313 non-null object\n", + "dtypes: object(9)\n", + "memory usage: 86.1+ MB\n" + ] + } + ], + "source": [ + "bioscan_df.info(show_counts = True)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "All images are labeled down to `order`, most (98.6%) are labeled to the `family` level, but we only have 22.5% and 7.5% with `genus` or `species` level designation, respectively." + ] + }, + { + "cell_type": "code", + "execution_count": 26, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "data_source 1\n", + "kingdom 1\n", + "phylum 1\n", + "class 1\n", + "order 19\n", + "family 494\n", + "genus 3444\n", + "species 8315\n", + "common 10635\n", + "dtype: int64" + ] + }, + "execution_count": 26, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "bioscan_df.nunique()" + ] + }, + { + "cell_type": "code", + "execution_count": 27, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "kingdom\n", + "Animalia 1128313\n", + "Name: count, dtype: int64" + ] + }, + "execution_count": 27, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "bioscan_df['kingdom'].value_counts()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "BIOSCAN is all `Animalia`." + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "Check we're not missing `family` designation when we have `genus`." + ] + }, + { + "cell_type": "code", + "execution_count": 28, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "Index: 254149 entries, 4401333 to 8403831\n", + "Data columns (total 9 columns):\n", + " # Column Non-Null Count Dtype \n", + "--- ------ -------------- ----- \n", + " 0 data_source 254149 non-null object\n", + " 1 kingdom 254149 non-null object\n", + " 2 phylum 254149 non-null object\n", + " 3 class 254149 non-null object\n", + " 4 order 254149 non-null object\n", + " 5 family 254149 non-null object\n", + " 6 genus 254149 non-null object\n", + " 7 species 84441 non-null object\n", + " 8 common 254149 non-null object\n", + "dtypes: object(9)\n", + "memory usage: 19.4+ MB\n" + ] + } + ], + "source": [ + "bioscan_df.loc[bioscan_df['genus'].notna()].info()" + ] + }, + { + "cell_type": "code", + "execution_count": 29, + "metadata": {}, + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
data_sourcekingdomphylumclassorderfamilygenusspeciescommon
5553984BIOSCANAnimaliaArthropodaInsectaDipteraPhoridaeMegaseliaNaNMegaselia
5115679BIOSCANAnimaliaArthropodaInsectaLepidopteraOecophoridaeIdioglossaidioglossa malaise5743Idioglossa idioglossa malaise5743
7746318BIOSCANAnimaliaArthropodaInsectaDipteraPhoridaeMegaseliaNaNMegaselia
5034837BIOSCANAnimaliaArthropodaInsectaDipteraChironomidaeChaetocladiuschaetocladius perennisChaetocladius chaetocladius perennis
6157434BIOSCANAnimaliaArthropodaInsectaDipteraPhoridaeApocephalusNaNAnt-decapitating Flies
7720398BIOSCANAnimaliaArthropodaInsectaDipteraPsychodidaePsychodapsychoda phalaenoidesPsychoda psychoda phalaenoides
6191559BIOSCANAnimaliaArthropodaInsectaDipteraPhoridaeMetopina_groupNaNMetopina_group
\n", + "
" + ], + "text/plain": [ + " data_source kingdom phylum class order family \\\n", + "5553984 BIOSCAN Animalia Arthropoda Insecta Diptera Phoridae \n", + "5115679 BIOSCAN Animalia Arthropoda Insecta Lepidoptera Oecophoridae \n", + "7746318 BIOSCAN Animalia Arthropoda Insecta Diptera Phoridae \n", + "5034837 BIOSCAN Animalia Arthropoda Insecta Diptera Chironomidae \n", + "6157434 BIOSCAN Animalia Arthropoda Insecta Diptera Phoridae \n", + "7720398 BIOSCAN Animalia Arthropoda Insecta Diptera Psychodidae \n", + "6191559 BIOSCAN Animalia Arthropoda Insecta Diptera Phoridae \n", + "\n", + " genus species \\\n", + "5553984 Megaselia NaN \n", + "5115679 Idioglossa idioglossa malaise5743 \n", + "7746318 Megaselia NaN \n", + "5034837 Chaetocladius chaetocladius perennis \n", + "6157434 Apocephalus NaN \n", + "7720398 Psychoda psychoda phalaenoides \n", + "6191559 Metopina_group NaN \n", + "\n", + " common \n", + "5553984 Megaselia \n", + "5115679 Idioglossa idioglossa malaise5743 \n", + "7746318 Megaselia \n", + "5034837 Chaetocladius chaetocladius perennis \n", + "6157434 Ant-decapitating Flies \n", + "7720398 Psychoda psychoda phalaenoides \n", + "6191559 Metopina_group " + ] + }, + "execution_count": 29, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "bioscan_df.loc[bioscan_df['genus'].notna()].sample(7)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "Observe that we do have instances where `common` is labeled as `Genus genus species`." + ] + }, + { + "cell_type": "code", + "execution_count": 30, + "metadata": {}, + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
data_sourcekingdomphylumclassorderfamilygenusspeciescommon
6710632BIOSCANAnimaliaArthropodaInsectaColeopteraScirtidaeNaNNaNAnimalia Arthropoda Insecta Coleoptera Scirtidae
4530081BIOSCANAnimaliaArthropodaInsectaHymenopteraMymaridaeNaNNaNAnimalia Arthropoda Insecta Hymenoptera Mymaridae
4368575BIOSCANAnimaliaArthropodaInsectaDipteraCecidomyiidaeNaNNaNAnimalia Arthropoda Insecta Diptera Cecidomyiidae
7790658BIOSCANAnimaliaArthropodaInsectaDipteraCecidomyiidaeNaNNaNAnimalia Arthropoda Insecta Diptera Cecidomyiidae
5208630BIOSCANAnimaliaArthropodaInsectaDipteraChironomidaeNaNNaNAnimalia Arthropoda Insecta Diptera Chironomidae
6199863BIOSCANAnimaliaArthropodaInsectaHymenopteraPlatygastridaeNaNNaNAnimalia Arthropoda Insecta Hymenoptera Platyg...
4882465BIOSCANAnimaliaArthropodaInsectaDipteraPsychodidaeNaNNaNAnimalia Arthropoda Insecta Diptera Psychodidae
\n", + "
" + ], + "text/plain": [ + " data_source kingdom phylum class order \\\n", + "6710632 BIOSCAN Animalia Arthropoda Insecta Coleoptera \n", + "4530081 BIOSCAN Animalia Arthropoda Insecta Hymenoptera \n", + "4368575 BIOSCAN Animalia Arthropoda Insecta Diptera \n", + "7790658 BIOSCAN Animalia Arthropoda Insecta Diptera \n", + "5208630 BIOSCAN Animalia Arthropoda Insecta Diptera \n", + "6199863 BIOSCAN Animalia Arthropoda Insecta Hymenoptera \n", + "4882465 BIOSCAN Animalia Arthropoda Insecta Diptera \n", + "\n", + " family genus species \\\n", + "6710632 Scirtidae NaN NaN \n", + "4530081 Mymaridae NaN NaN \n", + "4368575 Cecidomyiidae NaN NaN \n", + "7790658 Cecidomyiidae NaN NaN \n", + "5208630 Chironomidae NaN NaN \n", + "6199863 Platygastridae NaN NaN \n", + "4882465 Psychodidae NaN NaN \n", + "\n", + " common \n", + "6710632 Animalia Arthropoda Insecta Coleoptera Scirtidae \n", + "4530081 Animalia Arthropoda Insecta Hymenoptera Mymaridae \n", + "4368575 Animalia Arthropoda Insecta Diptera Cecidomyiidae \n", + "7790658 Animalia Arthropoda Insecta Diptera Cecidomyiidae \n", + "5208630 Animalia Arthropoda Insecta Diptera Chironomidae \n", + "6199863 Animalia Arthropoda Insecta Hymenoptera Platyg... \n", + "4882465 Animalia Arthropoda Insecta Diptera Psychodidae " + ] + }, + "execution_count": 30, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "bioscan_df.loc[bioscan_df['genus'].isna()].sample(7)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "Observe that when the `genus` is null, we get `common` of all higher order taxa available.\n", + "\n", + "This `common` assignment is something to fix. It could be part of the reason our first iteration did better with bugs (they got full taxa much more frequently)." + ] + }, + { + "cell_type": "code", + "execution_count": 31, + "metadata": {}, + "outputs": [ + { + "name": "stderr", + "output_type": "stream", + "text": [ + "/tmp/ipykernel_414099/2375049309.py:2: SettingWithCopyWarning: \n", + "A value is trying to be set on a copy of a slice from a DataFrame.\n", + "Try using .loc[row_indexer,col_indexer] = value instead\n", + "\n", + "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n", + " bioscan_df['duplicate'] = bioscan_df.duplicated(subset = taxa, keep = 'first')\n" + ] + } + ], + "source": [ + "#number of unique 7-tuples in BIOSCAN\n", + "bioscan_df['duplicate'] = bioscan_df.duplicated(subset = taxa, keep = 'first')\n", + "bioscan_df_unique_taxa = bioscan_df.loc[~bioscan_df['duplicate']]" + ] + }, + { + "cell_type": "code", + "execution_count": 32, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "Index: 10635 entries, 4319650 to 8403725\n", + "Data columns (total 10 columns):\n", + " # Column Non-Null Count Dtype \n", + "--- ------ -------------- ----- \n", + " 0 data_source 10635 non-null object\n", + " 1 kingdom 10635 non-null object\n", + " 2 phylum 10635 non-null object\n", + " 3 class 10635 non-null object\n", + " 4 order 10635 non-null object\n", + " 5 family 10621 non-null object\n", + " 6 genus 10203 non-null object\n", + " 7 species 8316 non-null object\n", + " 8 common 10635 non-null object\n", + " 9 duplicate 10635 non-null bool \n", + "dtypes: bool(1), object(9)\n", + "memory usage: 841.2+ KB\n" + ] + } + ], + "source": [ + "bioscan_df_unique_taxa.info(show_counts = True)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "Our number of unique 7-tuple taxa corresponds with the number of unique values for `common`, interesting, yet not surpising considering the way that `common` seems to have been generated. The actual common names that are in that column likley aren't repeated across species.\n", + "\n", + "We should be able to fill all in for all values of `species` that also have `genus` indicated. Is `genus` labeled for all entries with `species` labeled? " + ] + }, + { + "cell_type": "code", + "execution_count": 33, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "Index: 84450 entries, 4401381 to 8403808\n", + "Data columns (total 10 columns):\n", + " # Column Non-Null Count Dtype \n", + "--- ------ -------------- ----- \n", + " 0 data_source 84450 non-null object\n", + " 1 kingdom 84450 non-null object\n", + " 2 phylum 84450 non-null object\n", + " 3 class 84450 non-null object\n", + " 4 order 84450 non-null object\n", + " 5 family 84450 non-null object\n", + " 6 genus 84441 non-null object\n", + " 7 species 84450 non-null object\n", + " 8 common 84450 non-null object\n", + " 9 duplicate 84450 non-null bool \n", + "dtypes: bool(1), object(9)\n", + "memory usage: 6.5+ MB\n" + ] + } + ], + "source": [ + "bioscan_df.loc[bioscan_df.species.notna()].info(show_counts = True)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "There are only 9 images where the `species` is labeled, but the `genus` isn't." + ] + }, + { + "cell_type": "code", + "execution_count": 34, + "metadata": {}, + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
data_sourcekingdomphylumclassorderfamilygenusspeciescommonduplicate
4406083BIOSCANAnimaliaArthropodaInsectaHymenopteraEulophidaeNaNtetramalaise01 malaise6997tetramalaise01 malaise6997False
4426320BIOSCANAnimaliaArthropodaInsectaHymenopteraEulophidaeNaNtetramalaise01 malaise7245tetramalaise01 malaise7245False
4438796BIOSCANAnimaliaArthropodaInsectaHymenopteraEulophidaeNaNtetramalaise01 malaise7245tetramalaise01 malaise7245True
4450980BIOSCANAnimaliaArthropodaInsectaHymenopteraEulophidaeNaNtetramalaise01 malaise4739tetramalaise01 malaise4739False
4459021BIOSCANAnimaliaArthropodaInsectaHymenopteraEulophidaeNaNtetramalaise01 malaise7245tetramalaise01 malaise7245True
6196384BIOSCANAnimaliaArthropodaInsectaHymenopteraEulophidaeNaNtetramalaise01 malaise7544tetramalaise01 malaise7544False
7322471BIOSCANAnimaliaArthropodaInsectaHymenopteraEulophidaeNaNtetramalaise01 malaise4749tetramalaise01 malaise4749False
7766002BIOSCANAnimaliaArthropodaInsectaHymenopteraEulophidaeNaNtetramalaise01 malaise4739tetramalaise01 malaise4739True
8292691BIOSCANAnimaliaArthropodaInsectaHymenopteraEulophidaeNaNtetramalaise01 malaise4739tetramalaise01 malaise4739True
\n", + "
" + ], + "text/plain": [ + " data_source kingdom phylum class order family \\\n", + "4406083 BIOSCAN Animalia Arthropoda Insecta Hymenoptera Eulophidae \n", + "4426320 BIOSCAN Animalia Arthropoda Insecta Hymenoptera Eulophidae \n", + "4438796 BIOSCAN Animalia Arthropoda Insecta Hymenoptera Eulophidae \n", + "4450980 BIOSCAN Animalia Arthropoda Insecta Hymenoptera Eulophidae \n", + "4459021 BIOSCAN Animalia Arthropoda Insecta Hymenoptera Eulophidae \n", + "6196384 BIOSCAN Animalia Arthropoda Insecta Hymenoptera Eulophidae \n", + "7322471 BIOSCAN Animalia Arthropoda Insecta Hymenoptera Eulophidae \n", + "7766002 BIOSCAN Animalia Arthropoda Insecta Hymenoptera Eulophidae \n", + "8292691 BIOSCAN Animalia Arthropoda Insecta Hymenoptera Eulophidae \n", + "\n", + " genus species common \\\n", + "4406083 NaN tetramalaise01 malaise6997 tetramalaise01 malaise6997 \n", + "4426320 NaN tetramalaise01 malaise7245 tetramalaise01 malaise7245 \n", + "4438796 NaN tetramalaise01 malaise7245 tetramalaise01 malaise7245 \n", + "4450980 NaN tetramalaise01 malaise4739 tetramalaise01 malaise4739 \n", + "4459021 NaN tetramalaise01 malaise7245 tetramalaise01 malaise7245 \n", + "6196384 NaN tetramalaise01 malaise7544 tetramalaise01 malaise7544 \n", + "7322471 NaN tetramalaise01 malaise4749 tetramalaise01 malaise4749 \n", + "7766002 NaN tetramalaise01 malaise4739 tetramalaise01 malaise4739 \n", + "8292691 NaN tetramalaise01 malaise4739 tetramalaise01 malaise4739 \n", + "\n", + " duplicate \n", + "4406083 False \n", + "4426320 False \n", + "4438796 True \n", + "4450980 False \n", + "4459021 True \n", + "6196384 False \n", + "7322471 False \n", + "7766002 True \n", + "8292691 True " + ] + }, + "execution_count": 34, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "bioscan_df.loc[(bioscan_df.species.notna()) & (bioscan_df.genus.isna())]" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "It seems the `genus` is `tetramalaise01` ([genus page for tetraMalaise01](https://v3.boldsystems.org/index.php/Taxbrowser_Taxonpage?taxid=1074204))." + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "In general, when the species is listed in BIOSCAN it is listed as `genus-species`. This is certainly true here and would work to fill the 9 missing genera." + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "### EOL" + ] + }, + { + "cell_type": "code", + "execution_count": 35, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "Index: 6621365 entries, 0 to 10436520\n", + "Data columns (total 9 columns):\n", + " # Column Non-Null Count Dtype \n", + "--- ------ -------------- ----- \n", + " 0 data_source 6621365 non-null object\n", + " 1 kingdom 3919403 non-null object\n", + " 2 phylum 3917533 non-null object\n", + " 3 class 2842328 non-null object\n", + " 4 order 3875193 non-null object\n", + " 5 family 3906994 non-null object\n", + " 6 genus 5119837 non-null object\n", + " 7 species 4408570 non-null object\n", + " 8 common 6621365 non-null object\n", + "dtypes: object(9)\n", + "memory usage: 505.2+ MB\n" + ] + } + ], + "source": [ + "eol_df.info(show_counts = True)" + ] + }, + { + "cell_type": "code", + "execution_count": 36, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "data_source 1\n", + "kingdom 3\n", + "phylum 49\n", + "class 119\n", + "order 679\n", + "family 5446\n", + "genus 89115\n", + "species 180738\n", + "common 518118\n", + "dtype: int64" + ] + }, + "execution_count": 36, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "eol_df.nunique()" + ] + }, + { + "cell_type": "code", + "execution_count": 45, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "data_source 1\n", + "kingdom 3\n", + "phylum 31\n", + "class 66\n", + "order 284\n", + "family 1480\n", + "genus 48126\n", + "species 0\n", + "common 148959\n", + "duplicate 2\n", + "dtype: int64" + ] + }, + "execution_count": 45, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "eol_df.loc[eol_df.species.isna()].nunique()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "There are 570,515 unique page IDs from EOL, which clearly represent varying levels of taxa.\n", + "\n", + "Unique species + unique common where species is null does not reach this number." + ] + }, + { + "cell_type": "code", + "execution_count": 37, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "kingdom\n", + "Metazoa 2207522\n", + "Archaeplastida 1481804\n", + "Fungi 230077\n", + "Name: count, dtype: int64" + ] + }, + "execution_count": 37, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "eol_df['kingdom'].value_counts()" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "EOL uses `Metazoa` and `Archaeplastida` in place of `Animalia` and `Plantae`, respectively. These designations will need to be merged.\n", + "\n", + "We have already observed that not all ranks are filled in at the higher levels, sometimes having just one gap. It seems this is particularly common for `class` (it has the least non-null values of any taxa by far)." + ] + }, + { + "cell_type": "code", + "execution_count": 38, + "metadata": {}, + "outputs": [ + { + "name": "stderr", + "output_type": "stream", + "text": [ + "/tmp/ipykernel_414099/52183210.py:2: SettingWithCopyWarning: \n", + "A value is trying to be set on a copy of a slice from a DataFrame.\n", + "Try using .loc[row_indexer,col_indexer] = value instead\n", + "\n", + "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n", + " eol_df['duplicate'] = eol_df.duplicated(subset = taxa, keep = 'first')\n" + ] + } + ], + "source": [ + "#number of unique 7-tuples in EOL\n", + "eol_df['duplicate'] = eol_df.duplicated(subset = taxa, keep = 'first')\n", + "eol_df_unique_taxa = eol_df.loc[~eol_df['duplicate']]" + ] + }, + { + "cell_type": "code", + "execution_count": 39, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "Index: 422190 entries, 0 to 10436512\n", + "Data columns (total 10 columns):\n", + " # Column Non-Null Count Dtype \n", + "--- ------ -------------- ----- \n", + " 0 data_source 422190 non-null object\n", + " 1 kingdom 256471 non-null object\n", + " 2 phylum 256242 non-null object\n", + " 3 class 192666 non-null object\n", + " 4 order 253217 non-null object\n", + " 5 family 255792 non-null object\n", + " 6 genus 421118 non-null object\n", + " 7 species 372306 non-null object\n", + " 8 common 422190 non-null object\n", + " 9 duplicate 422190 non-null bool \n", + "dtypes: bool(1), object(9)\n", + "memory usage: 32.6+ MB\n" + ] + } + ], + "source": [ + "eol_df_unique_taxa.info(show_counts = True)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "The actual number of distinct creatures is likely higher, as evidenced below where we have unique common names listed in `common` but all taxa are null.\n", + "\n", + "We should be able to fill all in for all values of `species` that also have `genus` indicated. Is `genus` labeled for all entries with `species` labeled? " + ] + }, + { + "cell_type": "code", + "execution_count": 40, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "\n", + "Index: 4408570 entries, 1 to 10436520\n", + "Data columns (total 10 columns):\n", + " # Column Non-Null Count Dtype \n", + "--- ------ -------------- ----- \n", + " 0 data_source 4408570 non-null object\n", + " 1 kingdom 3562320 non-null object\n", + " 2 phylum 3560488 non-null object\n", + " 3 class 2496336 non-null object\n", + " 4 order 3526489 non-null object\n", + " 5 family 3555902 non-null object\n", + " 6 genus 4408326 non-null object\n", + " 7 species 4408570 non-null object\n", + " 8 common 4408570 non-null object\n", + " 9 duplicate 4408570 non-null bool \n", + "dtypes: bool(1), object(9)\n", + "memory usage: 340.6+ MB\n" + ] + } + ], + "source": [ + "eol_df.loc[eol_df.species.notna()].info(show_counts = True)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "It looks like we do have just under 250 images for which there's a `species` label, but no `genus` label. This is a sufficiently low number that we could go in to manually check them and hopefully match the the proper higher taxonomic ranks (considering some species names get reused across different genera)." + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "Let's get a quick sample of the `common` column for images both with and without `species` labels. " + ] + }, + { + "cell_type": "code", + "execution_count": 41, + "metadata": {}, + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
data_sourcekingdomphylumclassorderfamilygenusspeciescommonduplicate
7109927EOLArchaeplastidaTracheophytaNaNFabalesPolygalaceaePolygalaleendertziaePolygala leendertziaeTrue
1169608EOLFungiAscomycotaEurotiomycetesPyrenulalesPyrenulaceaePyrenuladermatodesPyrenula dermatodesTrue
2955438EOLArchaeplastidaTracheophytaNaNMagnolialesAnnonaceaeAsiminaobovatabigflower pawpawTrue
1551580EOLNaNNaNNaNNaNNaNNumeniushudsonicusNumenius hudsonicusTrue
3845702EOLNaNNaNNaNNaNNaNUrticagracilis holosericeastinging nettleTrue
3499959EOLMetazoaChordataGnathostomataGymnophionaIchthyophiidaeIchthyophisbombayensisBombay CaecilianTrue
4155103EOLFungiBasidiomycotaAgaricomycetesBoletalesSuillaceaeSuillusplacidusSlippery white boleteTrue
\n", + "
" + ], + "text/plain": [ + " data_source kingdom phylum class \\\n", + "7109927 EOL Archaeplastida Tracheophyta NaN \n", + "1169608 EOL Fungi Ascomycota Eurotiomycetes \n", + "2955438 EOL Archaeplastida Tracheophyta NaN \n", + "1551580 EOL NaN NaN NaN \n", + "3845702 EOL NaN NaN NaN \n", + "3499959 EOL Metazoa Chordata Gnathostomata \n", + "4155103 EOL Fungi Basidiomycota Agaricomycetes \n", + "\n", + " order family genus species \\\n", + "7109927 Fabales Polygalaceae Polygala leendertziae \n", + "1169608 Pyrenulales Pyrenulaceae Pyrenula dermatodes \n", + "2955438 Magnoliales Annonaceae Asimina obovata \n", + "1551580 NaN NaN Numenius hudsonicus \n", + "3845702 NaN NaN Urtica gracilis holosericea \n", + "3499959 Gymnophiona Ichthyophiidae Ichthyophis bombayensis \n", + "4155103 Boletales Suillaceae Suillus placidus \n", + "\n", + " common duplicate \n", + "7109927 Polygala leendertziae True \n", + "1169608 Pyrenula dermatodes True \n", + "2955438 bigflower pawpaw True \n", + "1551580 Numenius hudsonicus True \n", + "3845702 stinging nettle True \n", + "3499959 Bombay Caecilian True \n", + "4155103 Slippery white bolete True " + ] + }, + "execution_count": 41, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "# existing species label\n", + "eol_df.loc[eol_df.species.notna()].sample(7)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "Good, some of these do have common names. We could check numbers of `common` where it doesn't match the `genus-species` form for a better count, but our random sample is promising." + ] + }, + { + "cell_type": "code", + "execution_count": 42, + "metadata": {}, + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
data_sourcekingdomphylumclassorderfamilygenusspeciescommonduplicate
2635269EOLNaNNaNNaNNaNNaNNaNNaNtārangaTrue
4404459EOLNaNNaNNaNNaNNaNNaNNaNTrans-Pecos thimbleheadTrue
575560EOLNaNNaNNaNNaNNaNNaNNaNCremastobaeusTrue
456164EOLNaNNaNNaNNaNNaNNaNNaNAmerican red raspberryTrue
10332391EOLNaNNaNNaNNaNNaNNaNNaNStraight-lined Vaxi MothTrue
6082438EOLNaNNaNNaNNaNNaNNaNNaNRock ptarmiganTrue
3443067EOLNaNNaNNaNNaNNaNNaNNaNCatapion meieriTrue
\n", + "
" + ], + "text/plain": [ + " data_source kingdom phylum class order family genus species \\\n", + "2635269 EOL NaN NaN NaN NaN NaN NaN NaN \n", + "4404459 EOL NaN NaN NaN NaN NaN NaN NaN \n", + "575560 EOL NaN NaN NaN NaN NaN NaN NaN \n", + "456164 EOL NaN NaN NaN NaN NaN NaN NaN \n", + "10332391 EOL NaN NaN NaN NaN NaN NaN NaN \n", + "6082438 EOL NaN NaN NaN NaN NaN NaN NaN \n", + "3443067 EOL NaN NaN NaN NaN NaN NaN NaN \n", + "\n", + " common duplicate \n", + "2635269 tāranga True \n", + "4404459 Trans-Pecos thimblehead True \n", + "575560 Cremastobaeus True \n", + "456164 American red raspberry True \n", + "10332391 Straight-lined Vaxi Moth True \n", + "6082438 Rock ptarmigan True \n", + "3443067 Catapion meieri True " + ] + }, + "execution_count": 42, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "# No species label\n", + "eol_df.loc[eol_df.species.isna()].sample(7)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "All `common` values are filled in with predominently common names. It seems we could map these back to some level of taxa with the taxon-common matching? Though why would these EOL images have common without any other designation?\n", + "\n", + "Good example of strange inconsistency: the `American red raspberry` is `Rubus strigosus`, and a quick Google search easily provides the entire taxonomy.\n", + "\n", + "`Cremastobaeus` seems to be a genus of wasp." + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "Checking for overlap between the three data sources will not give accurate results until these labels are standardized.\n", + "\n", + "In general, when the species is listed in BIOSCAN it is listed as `genus-species`, so we could generate a new column with just the species designation to try to get a more accurate count across species. \n", + "\n", + "For now, let's just take a quick look at genera across the datasets since they are more standardized (and listed more often in BIOSCAN)." + ] + }, + { + "cell_type": "code", + "execution_count": 46, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "there are 89115 genera in EOL\n", + "there are 4884 genera in inat21\n", + "there are 3444 genera in bioscan\n" + ] + } + ], + "source": [ + "eol_genera = list(eol_df.loc[eol_df['genus'].notna(), 'genus'].unique())\n", + "inat21_genera = list(inat21_df.loc[inat21_df['genus'].notna(), 'genus'].unique())\n", + "bioscan_genera = list(bioscan_df.loc[bioscan_df['genus'].notna(), 'genus'].unique())\n", + "\n", + "print(f\"there are {len(eol_genera)} genera in EOL\")\n", + "print(f\"there are {len(inat21_genera)} genera in inat21\")\n", + "print(f\"there are {len(bioscan_genera)} genera in bioscan\")" + ] + }, + { + "cell_type": "code", + "execution_count": 48, + "metadata": {}, + "outputs": [ + { + "name": "stdout", + "output_type": "stream", + "text": [ + "There are 4809 genera shared between EOL and iNat21.\n", + "There are 2720 genera shared between EOL and BIOSCAN.\n", + "There are 212 genera shared between iNat21 and BIOSCAN.\n", + "There are 212 genera shared between all three data sources.\n" + ] + } + ], + "source": [ + "gen_overlap = list(set(eol_genera) & set(inat21_genera))\n", + "print(f\"There are {len(gen_overlap)} genera shared between EOL and iNat21.\")\n", + "print(f\"There are {len(list(set(eol_genera) & set(bioscan_genera)))} genera shared between EOL and BIOSCAN.\")\n", + "print(f\"There are {len(list(set(inat21_genera) & set(bioscan_genera)))} genera shared between iNat21 and BIOSCAN.\")\n", + "print(f\"There are {len(list(set(gen_overlap) & set(bioscan_genera)))} genera shared between all three data sources.\")" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "BIOSCAN and iNat21's overlap of genera is completely contained in EOL." + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "## Overall Stats\n", + "\n", + "Keep in mind, this is without standardizing `kingdom` or fixing any other inconsistencies observed above." + ] + }, + { + "cell_type": "code", + "execution_count": 49, + "metadata": {}, + "outputs": [], + "source": [ + "import numpy as np" + ] + }, + { + "cell_type": "code", + "execution_count": 50, + "metadata": {}, + "outputs": [], + "source": [ + "avgs_all_images = []\n", + "std_all_images = []\n", + "avgs_labeled_images = []\n", + "std_labeled_images = []\n", + "for taxon in taxa_com[1:]: #taxa + common\n", + " num_taxon = df[taxon].nunique()\n", + " num_img_taxon = len(df.loc[df[taxon].notna()])\n", + " avg_all = 10436521/num_taxon\n", + " std_all = np.sqrt(10436521/num_taxon)\n", + " avg_labeled = num_img_taxon/num_taxon\n", + " std_labeled = np.sqrt(num_img_taxon/num_taxon)\n", + " avgs_all_images.append(avg_all)\n", + " std_all_images.append(std_all)\n", + " avgs_labeled_images.append(avg_labeled)\n", + " std_labeled_images.append(std_labeled)" + ] + }, + { + "cell_type": "code", + "execution_count": 51, + "metadata": {}, + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + "\n", + "\n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + " \n", + "
classaverage_all_imgsstandard_deviationavg_labeledstd_dev_labeled
0kingdom2.087304e+061444.7505671.546912e+061243.749090
1phylum2.129902e+05461.5086401.578100e+05397.253042
2class7.673912e+04277.0182764.895209e+04221.251188
3order1.362470e+04116.7248951.003962e+04100.197905
4family1.842281e+0342.9217991.360416e+0336.883823
5genus1.160723e+0210.7736848.965043e+019.468391
6species5.497362e+017.4144193.781941e+016.149748
7common1.979178e+014.4487951.979178e+014.448795
\n", + "
" + ], + "text/plain": [ + " class average_all_imgs standard_deviation avg_labeled \\\n", + "0 kingdom 2.087304e+06 1444.750567 1.546912e+06 \n", + "1 phylum 2.129902e+05 461.508640 1.578100e+05 \n", + "2 class 7.673912e+04 277.018276 4.895209e+04 \n", + "3 order 1.362470e+04 116.724895 1.003962e+04 \n", + "4 family 1.842281e+03 42.921799 1.360416e+03 \n", + "5 genus 1.160723e+02 10.773684 8.965043e+01 \n", + "6 species 5.497362e+01 7.414419 3.781941e+01 \n", + "7 common 1.979178e+01 4.448795 1.979178e+01 \n", + "\n", + " std_dev_labeled \n", + "0 1243.749090 \n", + "1 397.253042 \n", + "2 221.251188 \n", + "3 100.197905 \n", + "4 36.883823 \n", + "5 9.468391 \n", + "6 6.149748 \n", + "7 4.448795 " + ] + }, + "execution_count": 51, + "metadata": {}, + "output_type": "execute_result" + } + ], + "source": [ + "avg_std = pd.DataFrame(data = {'class': taxa_com[1:], 'average_all_imgs': avgs_all_images, 'standard_deviation': std_all_images,\n", + " 'avg_labeled': avgs_labeled_images, 'std_dev_labeled': std_labeled_images })\n", + "avg_std" + ] + }, + { + "cell_type": "code", + "execution_count": 52, + "metadata": {}, + "outputs": [], + "source": [ + "avg_std.to_csv(\"../data/avg_std_byClass.csv\", index = False)" + ] + }, + { + "cell_type": "markdown", + "metadata": {}, + "source": [ + "Observe that the Plant and Animal `kingdom`s are actually much more heavily represented than Fungi." + ] + }, + { + "cell_type": "code", + "execution_count": 54, + "metadata": {}, + "outputs": [], + "source": [ + "sns.set(rc = {'figure.figsize': (10,6)})" + ] + }, + { + "cell_type": "code", + "execution_count": 55, + "metadata": {}, + "outputs": [ + { + "data": { + "text/plain": [ + "" + ] + }, + "execution_count": 55, + "metadata": {}, + "output_type": "execute_result" + }, + { + "data": { + "image/png": "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", + "text/plain": [ + "
" + ] + }, + "metadata": {}, + "output_type": "display_data" + } + ], + "source": [ + "sns.histplot(df_taxa, x = 'kingdom')" + ] + }, + { + "cell_type": "code", + "execution_count": null, + "metadata": {}, + "outputs": [], + "source": [] + } + ], + "metadata": { + "kernelspec": { + "display_name": "venv", + "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": "3.8.10" + }, + "orig_nbformat": 4 + }, + "nbformat": 4, + "nbformat_minor": 2 +}