Preparing Real Campus Data with Overpass Turbo

Data prep code

Purpose

In this lab, you will download real building location data from OpenStreetMap (OSM) using Overpass Turbo, export it as CSV, and clean it using a provided C++ tool. This cleaned data will be used in later labs that focus on algorithm design, not wireless engineering.

This is not RF engineering. We intentionally ignore walls, interference, attenuation, and signal strength. Buildings are treated as simple geometric points. The goal is to practice algorithms on realistic, messy data.

What You Are Preparing

You are producing a file of demand points:

  • Each building → one point
  • Each point has coordinates in meters
  • Later labs will place Wi-Fi access points to cover these points

This preprocessing step is intentionally separated from the algorithmic work.

Step 1: Run an Overpass Turbo Query

Go to: https://overpass-turbo.eu

Paste the following query into the left panel:

[out:csv(::type,::id,name,building,"addr:housenumber","addr:street",::lat,::lon; true; ",")][timeout:90];
{{geocodeArea:Kenyon College}}->.a;
(
  way["building"](area.a);
  relation["building"](area.a);
);
out center;

Then click Run.

What this query does

  • Finds the OpenStreetMap area for Kenyon College
  • Extracts all objects tagged building=*
  • Outputs one row per building
  • out center; converts building polygons into a single point

Step 2: Export as CSV (Important)

Overpass Turbo does not provide a traditional “CSV download” button. Follow these steps exactly:

  1. After the query finishes, click the Data tab (top right)
  2. You should see output starting with: @type,@id,name,building,addr:housenumber,addr:street,@lat,@lon
  3. Select all text and copy
  4. Paste into a new file named: kenyon.csv

If your header does not include @lat and @lon, your query is incorrect.

Step 3: Clean and Normalize the Data

You will use a provided C++ tool called osm_wifi_prep (from the GitHub Classroom starter repository).

This tool:

  • Reads the raw Overpass CSV
  • Filters usable building rows
  • Converts latitude/longitude to a local meter-scale coordinate system

Build

make

Run

./osm_wifi_prep kenyon.csv kenyonout.csv

Output format

id,name,lat,lon,x_m,y_m
way:322703800,Mather Hall,40.3789,-82.3957,61.1248,454.442
...
  • lat, lon are original OSM values
  • x_m, y_m are coordinates in meters

Common Issues

“No usable rows found”

  • Cause: missing latitude/longitude columns
  • Fix: confirm out center; is in the query

Unnamed buildings

  • Many OSM buildings lack names
  • The tool uses the OSM ID as a fallback
  • This is acceptable for the lab

Extra buildings outside campus

  • OSM boundaries are imperfect
  • This realism is intentional

What You Should Observe

  • The data is real and messy
  • Input validation matters
  • Algorithm results depend on preprocessing choices

Keep kenyonout.csv. You will use it directly in the next assignment.

Scroll to Top