Exploring Divvy Bike Data
Fall 2022
HS 650 Data Science and Predictive Analysis
Course by Professor Ivo Dinov
Computational Medicine and Bioinformatics, University of Michigan
HS 650 Data Science and Predictive Analysis
Course by Professor Ivo Dinov
Computational Medicine and Bioinformatics, University of Michigan
Divvy Bikes, a bicycle sharing system owned by the Chicago Department of Transportation and have been an immensely successful all over Chicago and the ridership has risen significantly over the years. Through this data analytic study, I have attempted to analyze the Divvy Bike Trips Data acquired from the Kaggle and look for trends and patterns in ridership.
With these observable patterns I hypothesize that the trip duration is dependent on a few variables, which are if the user is a subscriber/customer/dependent, their gender, the temperature on that day, the day of the week and the overall weather on the day of the ride.
I use multiple linear regression to fit a model on the training data and predict the trip duration using this model on the test data. The results indicate that there is a statistical significance between trip duration and the input variables with the p-values being infinitesimally small and a good F-Statistic score. However, the model does not predict trip duration correctly and returns large errors. The project is completed with R.
I use multiple linear regression to fit a model on the training data and predict the trip duration using this model on the test data. The results indicate that there is a statistical significance between trip duration and the input variables with the p-values being infinitesimally small and a good F-Statistic score. However, the model does not predict trip duration correctly and returns large errors. The project is completed with R.
︎ Link to project on RPubs
︎ Link to code on GitHub
