The project had three goals:
- Predict TriMet ridership by bus line more accurately than current model
- Create an interface to create predictions of ridership in the future.
- Identify bus lines that are similar to one another
Data for the project was provided by TriMet and augmented with additional sources to describe 82 bus lines.
- 178 possible variables to predict and describe the 82 bus routes
- Variables on population, line characteristics, crime, and race
- Only about 4% of the data was missing and was imputed
- Jan-Mar 2016
- Skills Related:
- Data Munging, R, Shiny, GIS, Exploratory Analysis
- TriMet Ridership Projection App