Your first analysis

This page walks you through your first UNA analysis end-to-end — from a blank workspace file to a map of bus-stop accessibility for every building in Cambridge, MA. It takes about ten minutes.

By the end you will have:

  • A working UNA_Workspace.py in your local package folder,

  • A feather file and GeoJSON file with per-origin accessibility scores,

  • The result visualized in QGIS.

Prerequisites

Before you start, make sure you have:

  1. UNA installed — the una conda environment ready to activate. See Installation.

  2. The Boston tutorial data placed inside your project folder — the assumed layout is:

    my_project/              ← any folder (contains UNA_Workspace.py,
│                          copied from the repo's examples/ folder)
└── Boston/              ← tutorial data (ships in the repo at docs/Boston)

    The three files we’ll use are:

    • 20260703_PercLenNetwork_InnerCore.geojson — the Boston inner-core pedestrian network (69,957 edges).

    • Cambridge_building_centroids.geojson — every building in Cambridge (14,751 point centroids).

    • MA_bus_stops.geojson — 6,661 MBTA-region bus stops, with a weekly_departures column giving weekly departures per stop.

  3. QGIS installed — for opening and styling the output. Any version 3.44+ or 4.0.3+ works.

Step 1. Open the workspace file

Open Visual Studio Code and open your project folder. In the file tree on the left, click UNA_Workspace.py. This is where you edit UNA settings.

Open a Terminal panel in VS Code (Terminal → New Terminal) and activate the conda environment:

conda activate una

You’ll run every command from this terminal for the rest of the tutorial.

Step 2. Configure the analysis

Delete everything currently in UNA_Workspace.py and paste the following:

from urban_network_analysis import UNA

una = UNA()

una.settings.data_folder       = r"Boston"
una.settings.network_file      = "20260703_PercLenNetwork_InnerCore.geojson"
una.settings.origins_file      = "Cambridge_building_centroids.geojson"
una.settings.destinations_file = "MA_bus_stops.geojson"

una.settings.search_radius              = 500
una.settings.destination_weight_column  = "weekly_departures"

una.RunAccessibility()

That’s it — nine lines. Save the file (File → Save, or Ctrl+S / Cmd+S).

Here’s what each block does:

  • Lines 1–3 load UNA and create an una instance.

  • Lines 5–8 point at the three data layers (network + origins + destinations). r"Boston" assumes the Boston data folder sits inside your project folder; adjust the path if your layout differs.

  • Line 10 sets the search radius to 500 m — buildings can reach any bus stop within a 500-meter network walk.

  • Line 11 tells UNA to weight each bus stop by its weekly_departures column, so a stop with many weekly departures counts more than a stop with few.

  • Line 13 runs the accessibility analysis.

Step 3. Run it

In the VS Code terminal, run:

python UNA_Workspace.py

You should see log output that looks something like:

[UNA] Instance created.
[Network added] Network added from 20260703_PercLenNetwork_InnerCore.geojson, with 69957 edges.
[Origins added] Origins added from Cambridge_building_centroids.geojson, with 14751 points.
[Destinations added] Destinations added from MA_bus_stops.geojson, with 6661 points.
[UNA Accessibility] Using AccessibilityWElevation (no elevation — symmetric weights).
[Export Results] Combined results (with geometry) exported to Boston/Results/accessibility_2026-01-15_1042/Results.feather

The exact folder name will differ (it’s timestamped). The whole run takes about 30 seconds on a modern laptop.

Step 4. Find the output

Navigate to Boston/Results/accessibility_<timestamp>/ (the Results folder is created inside your data_folder). You should see:

  • Results.feather — the analysis output in Feather format.

  • Results.geojson — the same output in GeoJSON (easier to open in QGIS).

The GeoJSON has one point per origin (~20,000 buildings) with the following columns:

Column

Meaning

geometry

Building centroid location.

reach

Weighted count of bus stops within 500 m (sum of weekly_departures values).

gravity_ exponential

Gravity-weighted accessibility (exponential decay).

gravity_ logistic

Gravity-weighted accessibility (logistic decay).

knn_logistic

K-nearest-neighbor composite accessibility.

Step 5. Visualize in QGIS

Open QGIS. In the Browser panel, navigate to your Boston/Results/accessibility_<timestamp>/ folder and drag Results.geojson onto the map canvas.

Set the project CRS to match the data:

  • Project → Properties → CRS

  • Search 6491 and select EPSG:6491 (NAD83(2011) Massachusetts Mainland).

You should see 14,751 building centroids across Cambridge, all colored identically at first. To make the accessibility visible:

  1. Right-click the layer → Properties → Symbology.

  2. Change Single symbol to Graduated at the top.

  3. Set Value to reach.

  4. Click Classify, then choose a color ramp (Viridis is a good default).

  5. Click OK.

Now the map is colored by accessibility: dark points are buildings with low transit access, bright points are buildings with many daily bus departures within 500 m. You should see a clear pattern — Harvard Square, Central Square, and Kendall Square lit up, peripheral neighborhoods dark.

To load the network as a background:

  • Drag 20260703_PercLenNetwork_InnerCore.geojson (from the Boston/ folder, not from Results) onto the canvas.

  • Style it with a thin gray line.

You now have a real accessibility map showing how many weekly bus departures are within a 500m walkshed from each address point in Cambridge, MA.

Reach to bus stops 500m

Reach to bus stops in Cambridge, MA in a 500m walkshed (weighted by weekly departures at each stop)

What you built

This tutorial ran the simplest possible accessibility analysis: Reach (bus stops within 500 m, weighted by daily departures) plus three companion metrics that came for free. The pattern you’d see in the map — central density, peripheral scarcity — is the empirical signature every transit-accessibility analysis in every city produces.

To go deeper, try modifying the workspace file and re-running. Some small experiments:

  • Change search_radius to 800 and rerun. How does the pattern spread outward?

  • Set una.settings.use_nearest_destination = True. Now each building’s accessibility depends only on its single nearest stop. How does the map change?

  • Set una.settings.elevation = True and rerun. Buildings on Toomemägi (the central hill) should get slightly lower scores.

Next steps