Finding your way through the supermarket - the easy way. 🤙
The project provides two features:
- We can calculate the shortest path between any two items, using the A* search algorithm.
- We can calculate the shortest tour from entrance to checkout that passes all items on your shopping list, using the Concorde TSP solver.
The features are exposed through an OpenAPI v3 API and made available in a Docker container.
For end-users, we also developed an Android app. You can see a work in progress version here:
There exist many approaches for indoor positioning, mostly using wireless technologies like Wi-Fi and Bluetooth. However, all approaches come with an expensive setup, that involves purchasing expensive hardware and hiring qualified staff to perform the roll-out. This makes existing systems unfeasible for a project, that would invovle tenthousands of stores.
Instead, our approach requires no hardware and is easy to roll out: QR codes! We position QR at critical locations inside the store, each containing information about its location. The QR code could be positioned on a poster, explaining how to use this new feature.
As part of this hackathon, we developed the nx-concorde package, a Python package for solving the Travelling Salesperson Problem (TSP) for graphs. We found it very useful, so we decided to make it open source.
Disclaimer: The work on this package started prior to the hackathon weekend, as you can see in the commit history. Building a package like this over a single weekend surely wouldn't have been possible for us. We don't want to make our results seem more impressive than what we actually accomplished. This would be unfair towards the other participating teams and was also never intended this way. 🌈
Since our approach is API first, it would be simple to add additional logic to the navigation system, for example a recommender system for products that are already on the customers fastest path. This way we can systematically increase sales.
For more information about the API microservice, check the API README.
This project was put together over the course of a weekend during the Hackdays Baden-Württemberg hackathon. Our team consisted of:
- Hossam Waziry(Android Developer)
- Kiran Vasudev(Data Scientist)
- Jan-Benedikt Jagusch(Data Scientist)
See LICENSE for details. Please note that the underlying Concorde TSP solver is only available for academic research use. For other uses, contact William Cook for licensing options.