Transit Equity Scenario Analyzer (TESA)

An interactive Streamlit dashboard that layers TTC transit infrastructure over Toronto neighbourhood census data. Built for the Transit Data Challenge 2026.

The goal is to make it easy to see which communities are most transit-dependent, where service gaps exist, and where a new station would do the most good — and to let anyone sketch a proposed extension and immediately see its equity impact.

Hosted URL: https://transitequityscenarioanalyzer.streamlit.app/

---

Running Locally

Prerequisites

You'll need Python 3.10 or newer. If you're not sure which version you have, open a terminal and run:

python --version

If you don't have Python, download it from python.org — make sure to check "Add Python to PATH" during installation on Windows.

Step 1 — Get the code

If you have Git installed:

git clone <repo-url>
cd tesa

Or download and unzip the repository from GitHub, then open a terminal in the unzipped folder.

Step 2 — Create a virtual environment (recommended)

A virtual environment keeps the project's dependencies isolated from the rest of your system.

Mac / Linux:

python -m venv venv
source venv/bin/activate

Windows (Command Prompt):

python -m venv venv
venv\Scripts\activate

Windows (PowerShell):

python -m venv venv
venv\Scripts\Activate.ps1

You'll see (venv) at the start of your terminal prompt when the environment is active. Run the activate command again any time you open a new terminal.

Step 3 — Install dependencies

pip install -r requirements.txt

This installs Streamlit, GeoPandas, Folium, and the other libraries the app needs. It may take a minute or two on first run.

Windows note: GeoPandas can be tricky to install on Windows. If pip install -r requirements.txt fails, try installing via conda instead:

conda install geopandas folium -c conda-forge
pip install streamlit streamlit-folium openpyxl

Step 4 — Add the data files

The app expects a data/ folder in the same directory as app.py. By default, a zip file is added to this repo so you can see 2021 data. When zipped, the folder contains:

File	Where to get it
data/census_data.xlsx	Statistics Canada 2021 census metrics by Toronto neighbourhood
data/Neighbourhoods.geojson	City of Toronto Open Data portal — search "Neighbourhoods"
data/routes.txt data/trips.txt data/stops.txt data/shapes.txt data/stop_times.txt	TTC GTFS feed — download from the TTC open data page and extract the zip directly into data/
data/ttc-subway-shapefile-wgs84.zip	TTC subway shapefile — optional; only needed if you want to replace the built-in Line 3 geometry
data/stations_precomputed.csv	Generated by running preprocess_gtfs.py locally — already committed to this repo
data/stops_classified.csv	Generated by running preprocess_gtfs.py locally — already committed to this repo

If a file is missing, a red dot appears next to it in the sidebar. You can also upload replacement files directly through the sidebar UI without restarting the app.

Step 5 — (Re-)generate pre-computed data files if needed

The two CSV files above (stations_precomputed.csv and stops_classified.csv) are already committed to the repo, so you only need this step if you've updated the GTFS feed.

Run the following script locally — it reads stop_times.txt once and writes the two lightweight CSVs that the app uses at runtime:

python preprocess_gtfs.py

Then commit the updated CSVs before redeploying.

Why this exists: stop_times.txt can be several hundred MB. Streamlit Cloud's 1 GB memory limit means the app cannot safely scan the full file at startup. Pre-processing it locally and committing the results keeps the hosted app fast and stable.

Step 6 — Run the app

streamlit run app.py

The app will open automatically in your browser at http://localhost:8501. The first load takes a few extra seconds while it parses the GTFS feed — subsequent loads are cached.

---

What's in each tab

Census Map

Choropleth of any of 12 census metrics — transit use, commute duration, departure time bands, income, low-income rate, shelter cost burden, renters, seniors, recent immigrants, and visible minority share. Subway lines render in official TTC colours (yellow / green / blue / purple); bus and streetcar overlays are toggleable from the sidebar. Hover any neighbourhood for its value and underlying count.

Census Analysis

Ranked top/bottom-10 tables for each equity-relevant metric, a distribution bar chart across all 158 neighbourhoods, and a per-neighbourhood profile lookup that shows every indicator at once. Useful for spotting where multiple disadvantages overlap.

Equity Gap

Brings together two computed scores to surface underserved communities:

1. Transit Access Score — counts stops and stations of each mode within walking distance of each neighbourhood, applies mode weights (local bus ×1.0, express ×1.5, streetcar ×1.5, subway ×3.0), divides by neighbourhood area (km²), and normalizes to 0–100.
2. Need Score — a weighted composite of census vulnerability indicators (low income, seniors, renters, shelter burden, visible minority share, etc.), each normalized 0–100 and combined into a single score. Weights are adjustable via sliders.
3. Equity Gap = Need Score − Transit Access Score. A positive gap means a neighbourhood has greater need than its current level of service.

The tab shows an interactive choropleth, a top-15 most-underserved table, and bar charts for each component.

Extension Simulator

A freehand drawing tool layered over the census choropleth. Use the toolbar to:

• Polyline tool — draw a proposed route alignment
• Marker tool — place a marker at each proposed station
• Circle tool — sketch ~500 m walking catchment zones around stations
• Measure tool — estimate route distances

When you're done, click Export GeoJSON to download your proposal as subway_proposal.geojson. Drawings are not saved between sessions, so export before you close the tab.

Simulation Comparison

Upload the GeoJSON exported from the Extension Simulator. The app adds your proposed stations to the existing network, re-computes the Transit Access Score, recalculates the Equity Gap, and compares results side-by-side against the baseline. You'll see:

• Impact summary cards (stations added, neighbourhoods affected, average change in access and equity gap)
• Side-by-side baseline vs. simulated choropleth maps
• A delta bar chart showing which neighbourhoods improved most
• A full comparison table for all 158 neighbourhoods

About

Data source table, privacy compliance note, full metric definitions, and methodology notes.

---

Data sources

Dataset	Source	Licence
Census data	Statistics Canada, 2021	Open Government Licence
Neighbourhood boundaries	City of Toronto Open Data	Open Government Licence
GTFS feed	Toronto Transit Commission	Open Data
Subway shapefile	TTC / OpenStreetMap	Open Data

All datasets are open/public with no personally identifiable information. No individual travel records, smartcard data, or device identifiers are used. This project complies with the Transit Data Challenge Privacy-First Principle.

---

Technical notes

• stop_times.txt is not read at runtime. Subway station data and stop classifications are pre-processed locally using preprocess_gtfs.py and committed to the repo as data/stations_precomputed.csv and data/stops_classified.csv. This keeps the hosted app within Streamlit Cloud's memory limits. If you update the GTFS feed, re-run preprocess_gtfs.py locally and commit the regenerated CSV files before redeploying.
• Neighbourhood name matching between the census and GeoJSON is done by string equality after a small set of known corrections. About six neighbourhoods have residual naming mismatches and will appear unshaded on the choropleth.
• Percentage values in the app represent each neighbourhood's share of the city-wide total for that metric, not the neighbourhood-internal rate. A value of 2.4% for "Public Transit %" means that neighbourhood accounts for 2.4% of all transit commuters in Toronto — not that 2.4% of that neighbourhood's residents take transit.
• Line 3 (Scarborough RT, closed 2023) is no longer in TTC's GTFS feed. The app uses hardcoded coordinates for its six stations and alignment.
• Simulation scores are normalized on the same baseline scale as the original Transit Access Score, so baseline and simulated values are directly comparable.

---

Developed by Anitra Roy, Maggie Wu, Sameha Tasnim · Last updated May 2026