Uber Driver Conversion Analysis

Inspiration

Although we understand that the task for this project is rather simpler compared to the other project options provided, we decided to challenger ourselves by thinking of ways to go above and beyond for this project option. Our inspiration stems from our desire to not only leverage the lessons we've learned from the Masters of Science in Business Analytics program, but also reach out of our comfort zones and try new techniques/methods foreign to us in order to create an exciting project from this simple topic.

What it does

A classification model using ensemble method (XG Boost and K Means Clustering) to determine if a sign up user will become a driver and what variables play a critical role in this classification process. Furthermore, our team devised a website that outlines the detail of our data analysis and machine learning models.

How we built it

We utilized different python libraries (NumPy, Pandas, Skicit-learn) to build the ensemble classification model and ran that model using the dataset provided by the competition. We build the node.js front end using java script and back end using Python related to the machine learning model.

Challenges we ran into

Since the dataset has plenty of missing values (more than 70% of rows contain at least a missing value), we had to be creative in choosing which classification model would best suit the nature of our data. From there, trial and error took place to find the optimal algorithm for the classification model and deeper model tuning is also involved. After that, pushing the machine learning model into the website was also challenging.

Accomplishments that we're proud of

Given that this project has a 24 hour duration, our team is very proud to complete the project to our desired expectation and standards! Proper teamwork and communication was very essential and we are glad we pulled through an "all nighter" together while remaining focused on the task at hand.

What we learned

We learned that building such a project requires more teamwork than anticipated. This is because one task is dependent on the completion of the previous task. In other words, one person can start working only if the other has completed the task. Because of this, each one of us had to perform proper time management and communication practices to maintain a steady work flow and navigate through the 24 hour tight deadline.