Week 8 — Mapping I & II

Foundations & Choropleths; Spatial Joins, Enrichment & Interactivity

📖 Background & Motivation

This two‑session module builds a solid foundation for geographic data visualization and then applies it to real analytical workflows. In Mapping I, students learn core geospatial concepts (data types, formats, and coordinate systems) and produce clear choropleth maps with appropriate normalization, classification, and legends. In Mapping II, students connect maps to real analysis tasks by performing spatial and attribute joins, enriching with socio‑economic context, and delivering interactive maps for exploration and communication.

🔎 Learning Objectives

  • Distinguish vector vs. raster data; recognize common geospatial file formats (Shapefile, GeoJSON, TopoJSON, GeoPackage).
  • Explain Coordinate Reference Systems (CRS) and projections; choose and document an appropriate CRS.
  • Create choropleths with sensible normalization, classification, color, and legends.
  • Perform spatial joins (point‑in‑polygon, polygon overlaps) and attribute joins on stable keys (e.g., FIPS/GEOID).
  • Enrich maps with socio‑economic variables (e.g., ACS) and build interactive web maps.

📚 Readings & Resources

🛠️ Setup

Install/verify:

pip install geopandas folium plotly mapclassify pyproj shapely pandas matplotlib

Notes:

  • If computing areas/lengths, reproject to an appropriate projected CRS (e.g., equal‑area).
  • Interactive HTML maps (Folium/Plotly) load tiles/scripts at view time.

🧭 Session 1 — Mapping I: Foundations & Choropleths (≈75 min)

1) Why map? When not to map

  • Use maps when spatial structure matters (clusters, corridors, gradients).
  • Prefer rates/ratios over raw counts; preview MAUP & ecological fallacy.

2) Geographic data fundamentals

  • Vector (points/lines/polygons/multiparts) vs Raster (continuous/categorical); resolution, extent.

3) File formats

  • Shapefile (.shp/.shx/.dbf/.prj) – common pitfalls: field name limits, encoding, missing .prj.
  • GeoJSON (WGS84 lon/lat); TopoJSON (shared arcs, smaller web payloads); GeoPackage, Parquet+WKB.

4) CRS & projections

  • Geographic vs Projected CRS; angular vs linear units.
  • WGS84 (EPSG:4326) vs Web Mercator (EPSG:3857) vs Equal‑Area (e.g., ESRI:102003 for CONUS).
  • Distortions (area/distance/direction); document CRS in captions.

5) Choropleths: principles

  • Appropriate use: polygon‑level rates with clear denominators and time windows.
  • Classification: Quantiles, Equal Interval, Natural Breaks (Jenks), Std‑Dev; 4–7 bins.
  • Color: sequential for ordered magnitudes; diverging for centered (±mean); avoid rainbow; accessible palettes.
  • Legends & captions: variable name + units, bin method, CRS, data period, source.

6) Demos (brief)

  • GeoPandas + Matplotlib: static choropleth; export PNG/SVG.
  • Plotly Express: interactive choropleth with hover; export HTML.
  • Folium: tile basemap, tooltips/popups, layer control; export HTML.
  • Download the Jupyter Notebook and Data

7) Mini‑exercise (10–15 min)

  • Given polygons + counts + population: make two choropleths (quantiles vs equal‑interval).
  • Add a 3‑sentence note comparing how binning changes the story.

🧭 Session 2 — Mapping II: Spatial Joins, Enrichment & Interactivity (≈75 min)

1) Spatial joins

  • Point‑in‑polygon (assign incidents/POIs to tracts), polygon‑to‑polygon (overlaps), nearest joins.
  • GeoPandas sjoin predicates: within, contains, intersects; geometry validity & CRS alignment.

2) Attribute joins & data hygiene

  • Join on stable keys (FIPS/GEOID). Handle types, leading zeros, whitespace, duplicates, missing keys.

3) Aggregation, normalization, and stability

  • Dissolve/aggregate to analysis geography; per‑capita / per‑household / per‑area rates.
  • Consider rate stabilization (mention: Empirical Bayes) for small populations.

4) Socio‑economic enrichment

  • Typical sources: ACS (income, education, housing, commute), CDC PLACES, BLS.
  • Temporal alignment and definitional consistency.

5) Interactive mapping patterns

  • Layer toggles (choropleth + points), marker clustering, heatmaps (when appropriate).
  • Filters (sliders/dropdowns), hover templates, annotations; geometry simplification & TopoJSON for performance.

6) Demos (brief)

  • Demo A: spatial join (points→polygons) → counts per polygon → normalized rate.
  • Demo B: attribute join with ACS; compute indicator; classify & map.
  • Demo C: interactive Folium (choropleth + proportional symbols, tooltips, layer control).
  • Demo D: interactive Plotly/Mapbox; export standalone HTML.

7) In‑class activity

  • Start‑to‑finish pipeline:
  • spatial join incidents → polygons,
  • enrich with ACS denominator,
  • build an interactive map with two layers,
  • 3–5 sentence reflection on design choices + limitations.
  • Download the Jupyter Notebook and Data

💻 Starter Code (snippets)

# GeoPandas read + quick plot
import geopandas as gpd
polys = gpd.read_file(gpd.datasets.get_path('naturalearth_lowres'))
polys.to_crs(3857).plot(edgecolor='white')

# Attribute join on a key
import pandas as pd
attrs = pd.read_csv('attrs.csv')  # contains GEOID and a value
joined = polys.merge(attrs, how='left', left_on='iso_a3', right_on='GEOID')

# Choropleth (static)
joined.plot(column='rate_per_10k', cmap='Blues', legend=True, edgecolor='0.7')

# Plotly Express (interactive)
import plotly.express as px
import json
geojson = json.loads(polys.to_json())
fig = px.choropleth(joined, geojson=geojson, locations=joined.index,
                    color='rate_per_10k', featureidkey='properties.index')
fig.update_geos(fitbounds='locations', visible=False)
fig.show()

# Spatial join (points → polygons)
pts = gpd.read_file('points.geojson').to_crs(polys.crs)
joined_pts = gpd.sjoin(pts, polys, predicate='within')
counts = joined_pts.groupby('polygon_id').size().rename('count')
result = polys.join(counts).fillna({'count':0})

🧪 In‑Class Activities (both sessions)

  • Work in pairs. Use provided notebook or your own dataset.
  • Maintain a caption block for each map (CRS, data period, binning, source, denominator).

🏠 Homework (Due next Tuesday, Oct 28)

  1. Pick a geographic unit (country/state/county/tract) + a topic.
  2. Build a reproducible pipeline (notebook) that includes:
  3. One static choropleth with clear legend + caption.
  4. One interactive map (Folium or Plotly) with at least two layers (e.g., choropleth + points) and tooltips.
  5. At least one join (spatial or attribute) and appropriate normalization.
  6. A 150–250 word reflection on design choices, denominators, classification, CRS, and limitations.
  7. Submit .ipynb and .html.

Rubric (10 pts)

  • Correct joins & CRS handling (3)
  • Sound normalization + clear classification/legend (3)
  • Interactive map quality (layers, tooltips, usability) (2)
  • Reflection clarity & critique of limitations (2)

✅ Submission Checklist

Before submitting, make sure:

  • Your assignment has fulfilled all the basic requirements listed above.

  • Use Quarto to render the notebook into HTML and zip the files for submission.

  • Double-check the visualizations and your reflections in the HTML are properly organized and displayed.