/moby-bikes

An end-to-end Data Science Project to optimize Operations team daily activities.

Primary LanguageJupyter NotebookMIT LicenseMIT

eBike Operations Optimization

Analytical Dashboard & Demand Forecasting

license

Web AppDocumentation

Moby bikes is an e-bike bike-share scheme in operation in Dublin. Electric powered bicycles may be rented from and returned to designated cycle stands inside the designated area.

As part of Moby Operations, there is a role called "eBike Operators" which among its responsibilities are distributing and relocating eBikes throughout the city, while performing safety checks and basic maintenance.

To optimize operations, we want to predict the demand for the next hours based on weather data in order to decide whether to increase fleet or is safe to perform safety checks and maintenance or to collect bikes for repairing.

In a nutshell (TL:DR)

  • Data Pipeline - Pulls data from APIs and dumps it into an AWS S3 bucket. Reads files from S3 and stores them in a MYSQL database.

  • Minimum Viable Product (MVP) - A web app displaying a few business metrics and charts.

  • Uses machine learning algorithms to predict the demand for the next hours based on weather data.

  • Moby Operations team will use Dashboard and Rental Demand Forecasting to make decisions and planning its daily operations.


Read full documentation