Random Data Project -- Austin Real-Time Road Conditions

I mentioned this dataset over here where I was playing around with Pandas and Polars.

Direct link to dataset https://data.austintexas.gov/Transportation-and-Mobility/Real-Time-Road-Conditions/ypbq-i42h/about_data

So, did I do anything with the data? Yes.

The dataset has a number of features, but I ended up playing around with using XGBoost to make a model that used air temp and relative humidity to predict the road grip rating (GOOD, FAIR, POOR) and using Open Meteo forecast data to feed into the model to get predictions. The model worked “ok” on GOOD, but the dataset is really unbalanced (with mostly GOOD ratings). The FAIR and POOR performance weren’t that great.

Ultimately though, you probably don’t need a computer to tell you to stay off the roads if it’s cold and wet outside. But wait, I did do something with the data. Looks like there’s currently 3 sensors reporting data. I made a dashboard to show you the current conditions.

https://austin-road-conditions.vercel.app

It’s raining outside as I type this, so standing water is believable.

repo: https://github.com/ThatOrJohn/austin-road-conditions