We found some that some taxi drivers may not sure where should he/she go to get a passenger. Also with the competitive jobs with Uber or Lyft drivers, they may less likely to get a job because nobody informs them.
What it does
We are using Clustering Technic, building a suggestion model for driver to drive to a better place to find a job. We also visualise the result once our suggestion system begins to work. Driver could easily access the App and find the best point for him.
How we built it
We are using the historical data from New York government data. More than 10 Million data are analysis. We used MongoDB to store all the information on Amazon AWS EC2. We also have a Flask Machine to apply our algorithm to calculating the realtime, real location result. A complete front end are created afterwards once we finished. Ref: http://www.nyc.gov/html/tlc/html/about/trip_record_data.shtml
Challenges we ran into
Anomaly in the data Big Data Challenge when store into the database Sometimes low performance of the suggestion system
Accomplishments that we're proud of
We have a live demo for our prediction! We have Simulation result We believe we can actually help taxi driver to find work!
What we learned
How to tackle with massive data How to visualise the data How to live predicting result How to Work Well in a Team
What's next for TaxiMatcher!
We will be keep pushing!