Spaces:
Sleeping
Sleeping
| # Generated from: Group_1.ipynb | |
| # Converted at: 2026-06-16T09:17:14.566Z | |
| # Next step (optional): refactor into modules & generate tests with RunCell | |
| # Quick start: pip install runcell | |
| # # GROUP 1 | |
| # Members: Lamis Alghamdi, Ayushi Uttamani, Annie Chien, Yu-Chen Tung | |
| import warnings | |
| warnings.filterwarnings("ignore") | |
| import streamlit as st | |
| st.set_page_config( | |
| page_title="Mesozoic Dinosaur Occurrences", | |
| page_icon="🦕", | |
| layout="wide", | |
| ) | |
| import pandas as pd | |
| import altair as alt | |
| from vega_datasets import data as vega_data | |
| import pycountry_convert as pc | |
| alt.data_transformers.disable_max_rows() | |
| df = pd.read_csv("dinosauria_occurrences.csv", low_memory=False) | |
| meso = df[(df["max_ma"] >= 66) & (df["max_ma"] <= 252)].copy() | |
| # We found out that some of the dinosaurs have mixed diet, so we need to simplify it. | |
| # | |
| def simplify_diet(d): | |
| if pd.isna(d): return "unknown" | |
| d = str(d).lower() | |
| if "carnivore" in d or "piscivore" in d or "insectivore" in d: return "carnivore" | |
| if "herbivore" in d: return "herbivore" | |
| if "omnivore" in d: return "omnivore" | |
| return "unknown" | |
| meso["diet_simple"] = meso["diet"].apply(simplify_diet) | |
| diet_colors = alt.Scale( | |
| domain=["carnivore", "herbivore", "omnivore", "unknown"], | |
| range=["#dc2626", "#16a34a", "#f59e0b", "#94a3b8"], | |
| ) | |
| # ### About the researcher dataset | |
| # | |
| # We wanted to visualize [scientific colonialism in paleontology](https://science.thewire.in/the-sciences/parachute-science-palaeontology-european-colonialism-indian-scientists/) by mapping where the fossil was found and where it is currently kept. The best way to do this is by using the longitude/latitude coordinates of the original location and the museum where it is currently housed. The museum column was only partially filled, and for the fossils where it was filled in, the museum country often matched the original location. So, the data was very biased and not useful for showing scientific colonialism in paleontology. | |
| # | |
| # We thought about using the [OpenAlex API](https://openalex.org/), which is a bibliographic catalog of scientific papers, authors, and institutions. The `reference_no` was used to look up each author's institutions and their country. We then extracted the first author’s country for each publication and merged it with the dataset we currently have. | |
| # | |
| # We want to do this because it enables a comparison between the geographic origin of fossil discoveries and the researchers studying them. Through this method, we can analyze global patterns of scientific contribution and identify potential imbalances in paleontological research across regions. | |
| enrich = pd.read_csv("references_enriched.csv") | |
| meso = meso.merge(enrich, on="reference_no", how="left") | |
| meso["cc"] = meso["cc"].replace({"UK": "GB"}) | |
| meso["first_author_country"] = meso["first_author_country"].replace({"UK": "GB"}) | |
| CONTINENT_MAP = {"AF":"Africa","AS":"Asia","EU":"Europe", | |
| "NA":"Americas","SA":"Americas","OC":"Oceania"} | |
| def cc_to_continent(cc): | |
| try: return CONTINENT_MAP.get(pc.country_alpha2_to_continent_code(cc)) | |
| except: return None | |
| meso["fossil_continent"] = meso["cc"].apply(cc_to_continent) | |
| meso["author_continent"] = meso["first_author_country"].apply(cc_to_continent) | |
| CONT_DOMAIN = ["Europe","Americas","Asia","Africa","Oceania"] | |
| CONT_RANGE = ["#e8c43d","#3b7ec1","#c95eb1","#3a8a3f","#d94f2c"] | |
| cont_scale = alt.Scale(domain=CONT_DOMAIN, range=CONT_RANGE) | |
| cont_sel = alt.selection_point(fields=["fossil_continent"], name="contSel", empty=True) | |
| # ## The period selector | |
| # | |
| # At first, we limited the users to choose the period, but we decided to increase the interactivity by allowing the users selecting the time range instead. | |
| # | |
| bins = list(range(66, 258, 5)) | |
| meso["ma_bin"] = pd.cut(meso["max_ma"], bins=bins, right=False).apply( | |
| lambda x: round((x.left + x.right) / 2, 1) if pd.notna(x) else None | |
| ) | |
| hist = ( | |
| meso.groupby("ma_bin", observed=False).size() | |
| .reset_index(name="count") | |
| .dropna(subset=["ma_bin"]) | |
| .sort_values("ma_bin") | |
| ) | |
| hist["ma_bin"] = hist["ma_bin"].astype(float) | |
| y_max = int(hist["count"].max()) | |
| boundary_df = pd.DataFrame({"ma": [201.3, 145.0]}) | |
| period_df = pd.DataFrame({ | |
| "ma": [228.0, 173.0, 105.0], | |
| "label": ["Triassic", "Jurassic", "Cretaceous"], | |
| "colour": ["#000000", "#000000", "#000000"], | |
| "y": [y_max * 1.06] * 3, | |
| }) | |
| brush = alt.selection_interval(encodings=["x"], name="time_brush", empty=True) | |
| x_enc = alt.X("ma_bin:Q", scale=alt.Scale(domain=[252, 66]), | |
| axis=alt.Axis(title="Million Years Ago (Ma)", tickCount=10, | |
| grid=True, gridDash=[4, 2])) | |
| y_enc = alt.Y("count:Q", scale=alt.Scale(domain=[0, y_max * 1.14]), | |
| axis=alt.Axis(title="Number of Occurrences", grid=True, gridDash=[4, 2])) | |
| area_base = (alt.Chart(hist) | |
| .mark_area(interpolate="monotone", color="#86efac", opacity=0.30, | |
| stroke="#16a34a", strokeWidth=1.5) | |
| .encode(x=x_enc, y=y_enc)) | |
| area_sel = (alt.Chart(hist) | |
| .mark_area(interpolate="monotone", color="#4ade80", opacity=0.65, | |
| stroke="#16a34a", strokeWidth=2.5) | |
| .encode(x=x_enc, y=y_enc) | |
| .transform_filter(brush)) | |
| rules = (alt.Chart(boundary_df) | |
| .mark_rule(color="#94a3b8", strokeDash=[5, 3], strokeWidth=1.3, opacity=0.8) | |
| .encode(x=alt.X("ma:Q", scale=alt.Scale(domain=[252, 66])))) | |
| period_text = (alt.Chart(period_df) | |
| .mark_text(fontSize=13, fontWeight="bold", align="center", font="Helvetica") | |
| .encode( | |
| x=alt.X("ma:Q", scale=alt.Scale(domain=[252, 66])), | |
| y=alt.Y("y:Q", scale=alt.Scale(domain=[0, y_max * 1.14])), | |
| text=alt.Text("label:N"), | |
| color=alt.Color("colour:N", scale=None), | |
| )) | |
| brush_chart = ( | |
| alt.layer(area_base, area_sel, rules, period_text) | |
| .properties(width=600, height=180, | |
| title=alt.TitleParams( | |
| "Fossil Diversity Over Time", | |
| subtitle="Drag horizontally to filter all panels below · 5 Ma bins", | |
| fontSize=18, font="Helvetica", fontWeight="bold", color="#000000", | |
| subtitleFont="Helvetica", subtitleFontSize=12, | |
| subtitleColor="#000000", anchor="start", offset=12, | |
| )) | |
| .add_params(brush) | |
| ) | |
| # ## The Continent filter | |
| # | |
| # We used the country_converter library to map countries to their continents. Users can select the continent(s) they want. | |
| # | |
| cont_df = pd.DataFrame({"fossil_continent": CONT_DOMAIN}) | |
| cont_buttons = ( | |
| alt.Chart(cont_df).mark_rect(cornerRadius=5) | |
| .encode( | |
| x=alt.X("fossil_continent:N", | |
| axis=alt.Axis(labelAngle=0, title=None, ticks=False, | |
| domain=False, labelFontSize=12, | |
| labelColor="#000000", labelFont="Helvetica", | |
| labelFontWeight="bold")), | |
| color=alt.condition( | |
| cont_sel, | |
| alt.Color("fossil_continent:N", scale=cont_scale, legend=None), | |
| alt.value("#e5e7eb")), | |
| tooltip=alt.Tooltip("fossil_continent:N", title="Click to filter"), | |
| ) | |
| .add_params(cont_sel) | |
| .properties( | |
| width=600, height=80, | |
| title=alt.TitleParams( | |
| "Choosing a Continent", | |
| subtitle="Click on a continent to choose it · hold SHIFT for multiple · click again to clear", | |
| fontSize=18, font="Helvetica", fontWeight="bold", color="#000000", | |
| subtitleFont="Helvetica", subtitleFontSize=12, subtitleFontWeight="bold", | |
| subtitleColor="#000000", anchor="start", offset=10, | |
| ) | |
| ) | |
| ) | |
| # ## The map | |
| # | |
| # Maps are important for this dataset. The dots represent different fossils, and the colors represent different diets. | |
| #: :generated: | |
| #: :model: Claude Opus 4.7 | |
| #: :prompt: Render a world map with country outlines and graticule using vega_datasets world_110m, and overlay fossil dig-site dots from `meso` colored by simplified diet. The dots should react to the time brush and continent pill selection. | |
| #: :changes: | |
| #: :response: | |
| countries = alt.topo_feature(vega_data.world_110m.url, "countries") | |
| base_world = alt.Chart(countries).mark_geoshape( | |
| fill="#f1f5f9", stroke="#94a3b8", strokeWidth=0.4 | |
| ).project(type="equirectangular") | |
| graticule = alt.Chart({"graticule": {"step": [30, 30]}}).mark_geoshape( | |
| fill=None, stroke="#cbd5e1", strokeWidth=0.3 | |
| ).project(type="equirectangular") | |
| paleo_points = ( | |
| alt.Chart(meso) | |
| .mark_circle(size=14, opacity=0.55, stroke="white", strokeWidth=0.3) | |
| .encode( | |
| longitude="lng:Q", | |
| latitude="lat:Q", | |
| color=alt.Color("diet_simple:N", scale=diet_colors, legend=None), | |
| tooltip=["accepted_name:N", "diet_simple:N", "max_ma:Q", "ma_bin:Q", "cc:N"], | |
| ) | |
| .project(type="equirectangular") | |
| .transform_filter(brush) | |
| .transform_filter(cont_sel) | |
| ) | |
| paleo_map = ( | |
| (base_world + graticule + paleo_points) | |
| .resolve_scale(color="independent") | |
| .properties( | |
| width=600, height=340, | |
| title=alt.TitleParams( | |
| "Where They're Found", | |
| subtitle="Modern dig sites · colored by diet", | |
| fontSize=18, font="Helvetica", fontWeight="bold", color="#000000", | |
| subtitleFont="Helvetica", subtitleFontSize=12, | |
| subtitleColor="#000000", anchor="start", offset=12, | |
| ) | |
| ) | |
| ) | |
| # ## KPIs | |
| #: :generated: | |
| #: :model: Claude Opus 4.7 | |
| #: :prompt: Build a 2x2 KPI panel for the dashboard with four metrics: distinct family count, total fossil occurrences, percentage of foreign-led papers (first-author country differs from fossil country), and percentage of rare species (accepted_name appearing fewer than RARITY_THRESHOLD times). Each KPI should respond to the time brush and continent selection. | |
| #: :changes: Improved the style. | |
| RARITY_THRESHOLD = 2 | |
| KPI_W, KPI_H = 200, 75 | |
| def kpi_title(text, subtitle): | |
| return alt.TitleParams( | |
| text, | |
| subtitle=subtitle, | |
| fontSize=14, font="Helvetica", fontWeight="bold", color="#000000", | |
| subtitleFont="Helvetica", subtitleFontSize=10, subtitleFontWeight="bold", | |
| subtitleColor="#000000", anchor="middle", | |
| ) | |
| # Families | |
| kpi_families = ( | |
| alt.Chart(meso.dropna(subset=["family"]).query("family != 'NO_FAMILY_SPECIFIED'")) | |
| .transform_filter(brush) | |
| .transform_filter(cont_sel) | |
| .transform_aggregate(n="distinct(family)") | |
| .mark_text(fontSize=36, fontWeight="bold", color="#16a34a", font="Helvetica", | |
| align="center", baseline="middle", dy=2) | |
| .encode(text=alt.Text("n:Q", format=",.0f")) | |
| .properties(width=KPI_W, height=KPI_H, | |
| title=kpi_title("Families", "Unique taxonomic families")) | |
| ) | |
| # Occurrences | |
| kpi_occurrences = ( | |
| alt.Chart(meso) | |
| .transform_filter(brush) | |
| .transform_filter(cont_sel) | |
| .transform_aggregate(n="count()") | |
| .mark_text(fontSize=36, fontWeight="bold", color="#3b7ec1", font="Helvetica", | |
| align="center", baseline="middle", dy=2) | |
| .encode(text=alt.Text("n:Q", format=",.0f")) | |
| .properties(width=KPI_W, height=KPI_H, | |
| title=kpi_title("Occurrences", "Total fossil records")) | |
| ) | |
| # Foreign-led | |
| meso_with_authors = meso.dropna(subset=["cc","first_author_country"]).copy() | |
| meso_with_authors["foreign_led"] = ( | |
| meso_with_authors["cc"] != meso_with_authors["first_author_country"] | |
| ).astype(int) | |
| kpi_foreign = ( | |
| alt.Chart(meso_with_authors) | |
| .transform_filter(brush) | |
| .transform_filter(cont_sel) | |
| .transform_aggregate(pct="mean(foreign_led)") | |
| .transform_calculate(pct_label='format(datum.pct * 100, ".0f") + "%"') | |
| .mark_text(fontSize=36, fontWeight="bold", color="#dc2626", font="Helvetica", | |
| align="center", baseline="middle", dy=2) | |
| .encode(text=alt.Text("pct_label:N")) | |
| .properties(width=KPI_W, height=KPI_H, | |
| title=kpi_title("Foreign-Led", "% papers with non-local first author")) | |
| ) | |
| # Rarity | |
| name_counts = meso["accepted_name"].value_counts() | |
| rare_names = set(name_counts[name_counts <= RARITY_THRESHOLD].index) | |
| meso["is_rare"] = meso["accepted_name"].isin(rare_names).astype(int) | |
| kpi_rarity = ( | |
| alt.Chart(meso) | |
| .transform_filter(brush) | |
| .transform_filter(cont_sel) | |
| .transform_aggregate(pct="mean(is_rare)") | |
| .transform_calculate(pct_label='format(datum.pct * 100, ".0f") + "%"') | |
| .mark_text(fontSize=36, fontWeight="bold", color="#f59e0b", font="Helvetica", | |
| align="center", baseline="middle", dy=2) | |
| .encode(text=alt.Text("pct_label:N")) | |
| .properties(width=KPI_W, height=KPI_H, | |
| title=kpi_title("Rare Species", f"% with ≤{RARITY_THRESHOLD} mentions")) | |
| ) | |
| kpi_panel = alt.vconcat( | |
| alt.hconcat(kpi_families, kpi_occurrences, spacing=8), | |
| alt.hconcat(kpi_foreign, kpi_rarity, spacing=8), | |
| spacing=8, | |
| ) | |
| #: :end: | |
| # ## The diet | |
| # Creating a chart that illustrates the dietary breakdown of known dinosaur genera | |
| # | |
| diet_bars = ( | |
| alt.Chart(meso) | |
| .mark_bar(stroke="white", strokeWidth=1) | |
| .encode( | |
| x=alt.X("count():Q", stack="normalize", | |
| axis=alt.Axis(title="Share", format="%")), | |
| y=alt.Y("dummy:N", axis=None, title=None), | |
| color=alt.Color("diet_simple:N", scale=diet_colors, | |
| legend=alt.Legend( | |
| title="Diet", orient="bottom", direction="horizontal", | |
| labelColor="#000000", titleColor="#000000", | |
| )), | |
| tooltip=["diet_simple:N", "count():Q"], | |
| ) | |
| .transform_calculate(dummy="'all'") | |
| .transform_filter(brush) | |
| .transform_filter(cont_sel) | |
| .properties( | |
| width=420, height=60, | |
| title=alt.TitleParams( | |
| "Diet Ratio", | |
| subtitle="Share within current selection", | |
| fontSize=18, font="Helvetica", fontWeight="bold", color="#000000", | |
| subtitleFont="Helvetica", subtitleFontSize=12, | |
| subtitleColor="#000000", anchor="start", offset=12, | |
| ) | |
| ) | |
| ) | |
| # ## Authors (Researchers) | |
| # | |
| # Plot the researchers' countries and where the fossil was found. | |
| df_r = meso.dropna(subset=["fossil_continent","author_continent"]) | |
| bubble_matrix = ( | |
| alt.Chart(df_r).mark_rect(stroke="white", strokeWidth=1) | |
| .encode( | |
| x=alt.X("fossil_continent:N", sort=CONT_DOMAIN, | |
| title="Fossil location", | |
| axis=alt.Axis(labelAngle=-30, labelFontSize=11, labelFont="Helvetica", | |
| labelColor="#000000", titleColor="#000000", | |
| titleFontSize=12, titleFont="Helvetica")), | |
| y=alt.Y("author_continent:N", sort=CONT_DOMAIN, | |
| title="Researcher", | |
| axis=alt.Axis(labelFontSize=11, labelFont="Helvetica", | |
| labelColor="#000000", titleColor="#000000", | |
| titleFontSize=12, titleFont="Helvetica")), | |
| color=alt.Color("count():Q", | |
| scale=alt.Scale(scheme="greens"), | |
| legend=alt.Legend(title="Occurrences", orient="right", | |
| labelColor="#000000", titleColor="#000000", | |
| labelFont="Helvetica", titleFont="Helvetica")), | |
| opacity=alt.condition(cont_sel, alt.value(1.0), alt.value(0.25)), | |
| tooltip=[ | |
| alt.Tooltip("fossil_continent:N", title="Fossil continent"), | |
| alt.Tooltip("author_continent:N", title="Researcher"), | |
| alt.Tooltip("count():Q", title="Occurrences"), | |
| ], | |
| ) | |
| .transform_filter(brush) | |
| .properties( | |
| width=420, height=240, | |
| title=alt.TitleParams( | |
| "Who Studies Whose Fossils?", | |
| subtitle="Researcher origin × fossil location", | |
| fontSize=18, font="Helvetica", fontWeight="bold", color="#000000", | |
| subtitleFont="Helvetica", subtitleFontSize=12, | |
| subtitleColor="#000000", anchor="start", offset=12, | |
| ) | |
| ) | |
| ) | |
| # ### The dashboard | |
| left_col = alt.vconcat(brush_chart, cont_buttons, paleo_map, spacing=20) | |
| right_col = alt.vconcat(kpi_panel, diet_bars, bubble_matrix, spacing=20) | |
| dashboard = ( | |
| alt.hconcat(left_col, right_col, spacing=24) | |
| .resolve_scale(color="independent") | |
| .properties( | |
| title=alt.TitleParams( | |
| "Mesozoic Dinosaur Occurrences", | |
| subtitle="Drag the time chart to filter · Click a continent pill to filter map, KPIs, and matrix", | |
| fontSize=22, font="Helvetica", fontWeight="bold", color="#000000", | |
| subtitleFont="Helvetica", subtitleFontSize=13, subtitleFontWeight="bold", | |
| subtitleColor="#000000", anchor="start", offset=12, | |
| ) | |
| ) | |
| .configure( | |
| background="white", | |
| padding={"left":40,"right":40,"top":30,"bottom":30}, | |
| ) | |
| .configure_view(strokeWidth=0) | |
| .configure_axis( | |
| labelFontSize=12, labelColor="#000000", labelFont="Helvetica", | |
| titleFontSize=12, titleColor="#000000", titleFont="Helvetica", | |
| titleFontWeight="bold", | |
| domainColor="#000000", tickColor="#000000", | |
| gridColor="#e5e7eb", | |
| ) | |
| .configure_legend( | |
| labelFontSize=12, titleFontSize=12, | |
| labelColor="#000000", titleColor="#000000", | |
| labelFont="Helvetica", titleFont="Helvetica", | |
| titleFontWeight="bold", padding=10, | |
| ) | |
| ) | |
| import warnings | |
| warnings.filterwarnings("ignore") | |
| st.title("Mesozoic Dinosaur Occurrences") | |
| st.markdown( | |
| "Interactive exploration of dinosaur fossil occurrences, diet, geography, and researcher origins." | |
| ) | |
| st.altair_chart( | |
| dashboard, | |
| use_container_width=True | |
| ) | |