Transit Guessr

Transit Guessr is a GeoGuessr-style game for subway systems.

It was initiall just NYC, but has been refactored to support multiple subway systems (currently Boston is the only other system, but more are on the way.)

There are essentially two apps running in this repo:

1. The Transit Guessr game
2. The analysis UI at /data: a visualization of 1 week of collected guess data. The data is no longer collected since it costs money to store and query, so it's just a one-time thing

Data

All station data is just manually created as a JSON file. For NYC I used NYC Open Data. For Boston, I used the publicly-available data and had the help of twitter user @2urbaneric1 to get the data into the correct format. That data is stored here

This is not an ideal data pipeline, but any sort of automated system felt like overkill.

For the analysis data, guess + final score info was pushed to firebase. That data was then aggregated with a bigquery query to generate src/data/guesses.tsx. An additional script was used to generate .geojson files from the firebase data for each individual station.

Again, this was not ideal and would require an annoying amount of manual effort to do again, but I don't really intend to do this specific analysis in an ongoing manner.

Development

This project was bootstrapped with Create React App.

You should be able to run it with npm run start

To run, you will need to create a Mapbox API token following these instructions. After creating, copy .env.local.example to .env.local and then have REACT_APP_MAPBOX_TOKEN map to your new token.

There are no tests, just manual testing.

The transit operator that the game is played with is controlled using the REACT_APP_TRANSITGUESSR_OPERATOR env var. Current valid values are mta and `mbta'

Contributing

Feel free to contibute any change! If you feel like making a contribution, some things I want to add are:

1. Different "levels" (hard = no parks, easy = shows streets)
2. More transit lines (Staten Island, LIRR, etc.)
3. Perhaps different transit agencies!
4. Any sorts of CSS or react improvments. These aren't my area of expertise so lots of the code here is pretty bad