Skip to content

Perform clustering to find patterns in rideshare data for Chicago and NYC

Notifications You must be signed in to change notification settings

s-saloni/RidesharePatterns

Repository files navigation

Mining for Rideshare Patterns

Team: Ronald Thompson, Jonathan Prindle, Saloni Sharma

Description

Using various data mining methods, primarily k-means clustering, we analyzed rideshare data from New York City and Chicago to discover patterns that can improve customer experiences and optimize profits for the rideshare drivers and companies.

Questions

As a curious bunch, we had many questions relating to our data. Here are a few:

  • At what times of day is there peak traffic?
  • Where are the most common locations where rideshares are utilized?
  • Do any patterns emerge that indicate when a driver are most likely to be tipped?
  • When and where can drivers maximize their profits? Where do customers pay the highest fares?

Results

We have gotten several results from our analyses to support our initial conjectures and answer our questions. The reason for which customers use rideshare determines the peak times. Notably, on weekdays, most rideshares take place before and after typical work hours. The most frequented locations for rideshares include the downtown area, city centers and airports.

The fares for rideshares generally support the idea that the larger the trip distance and the more passengers, the higher the fare per mile and the higher the likelihood of receiving a tip. However, our analysis showed that tipping does not fare well in the rideshare system. The majority of rideshares lack tips and on average, tips don't exceed eight dollars.

Applications

The analyses presented in this project can be a powerful tool for rideshare drivers and companies to increase their revenue. Combining the knowledge of peak traffic time and most popular locations can provide a strong starting point for drivers to establish a strategy for maintaining a steady flow of customers. In addition, with the results from our fare and tip analysis, drivers can increase the probability of receiving higher fares and tips. While the drivers achieve higher revenue by determining their customers needs in advance, customers will also benefit by receiving quick and easy rideshare services when they need them.

Links

View our final presentation video here.
Take a look at the final report here.

About

Perform clustering to find patterns in rideshare data for Chicago and NYC

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published