Bike Rental Data Extraction

Bike Rental Data Extraction

Tags
Data Extraction
R
Published
September 29, 2021
For my Bike Rental Prediction and Analysis Project, I used data that was compiled by Hadi Fanaee-T while they were at the University of Porto. In order to get a larger set of data, I decided to go to the source of their data [https://www.capitalbikeshare.com/system-data] and extract more years. To do this I downloaded the csv's and wrote a script to exactly match the outputs that Hadi originally conceived. I also had to find an alternative source of weather data since the original source that Hadi used no longer exists in the form it used to. For this, I used [https://www.visualcrossing.com/]. In order to match the weather situation which is not an exact measure of the weather, but a qualitative one, I decided to take the percentage of days that Hadi tagged as being either severe or mild weather and match that percentage to the worst precipitation days in my data set. While this doesn't work well as an exact replication of his process, it does get pretty close. In the future, I will remove this qualitative measure in favor of a more quantitative one, like total precipitation.