This project aims to explore bikeshare data from the Ford GoBike program across five cities in the San Francisco Bay Area. With the meteoric rise of bikeshare/scootershare services across the world, it was an interesting journey to revisit what kind of ridership patterns we could extract from bikeshare data. Following EDA, I also attempted to create a model that could predict ridership based on features extracted from the data; alas, the two years of data that we had is seemingly insufficient to create a robust model that does not overfit to the training set (not to mention, challenges in fitting to a time series dataset).
A writeup for this project can be found on Medium. It is the most concise overview of the project pipeline and findings; more detailed information about the project can be found in this repository (see below for details).
Data were collected from Kaggle; some data files (i.e. database.sqlite, status.csv) have been omitted from this repository because they were not used in the analysis, and they are too large to upload to GitHub.
Current data on the Ford GoBike program can be found here; this was not incorporated into this project.
Notebooks containing analysis can be found in notebooks, and data (raw and cleaned) can be found in data.
This project can be reproduced by creating a Python3 virtualenv with the dependencies listed in requirements.txt