Bike sharing services have an imbalance problem that's shared around the world. A simple search leads up to different studies that try to solve this problem in the best way that minimizes costs for the service manager and so the user's don't ever arrive to a full or empty station that can't be used.
This repository contains the code of the App Store app available for multiple cities around the world. All the data being used comes from the main repo linked below and shows users daily bike sharing demand predictions based on recent usage using custom neural networks trained specifically for each city.
- neural-bikes-cronjobs: Automated data upload & prediction generations
- neural-bikes: Python & Machine Learning process to generate neural network predictive models
- neural-bikes-ios-app
- neural-bikes-backend: Gathers data from InfluxDB database and uploads it to CloudKit
- neural-bikes-web: API & website
This app is completelly free and Open-Source if it's helped you can donate using Ko-Fi or using the In-App Purchases of the iOS app. There's no third party libraries involved or data shared with anyone.