The private used car sales industry is a minefield for everyone. Car service records are currently kept on outdated paper start books which are easily tampered with or lost, meaning that the car history is not fully known to the client. Buyers and sellers struggle to authenticate the health and value of cars, and often struggle to understand complicated technical information.

Our app delivers a predictive health report on any vehicle, notifies car buyers on how well the vehicle was maintained, and indicates exactly what needs servicing on the vehicle per manufacturers service interval specification. Blockchain was used for service logs as it uses immutable objects thus avoids any fraudulent behaviour.

The prototype was built as an IOS application interface without back end to demonstrate user interaction. A local python blockchain implementation was used to demonstrate the handling of vehicle service records in a semantic manner.

The vision of the final implementation: ERC20 utility token implementation of the prototype data structure.

The blockchain in particular proved to be a challenge, although the scope of the project was, at first, difficult to define. The vision encapsulates many aspects, not all of which we had time to develop fully. We finished with a prototype of the basic aspects which we hope to develop after the hackathon has finished!

Erika in particular took on the challenge of understanding blockchain and implementing our ideas in python which was no mean feat. As a group, we kept great communication throughout the project, coming up with new ideas and consistently improving our vision.

Erika learned how to implement blockchain! As a group we learned how to zone in on a vision; re-evaluating weak ideas and improving them. Everyone in the team learned new skills which will be invaluable going forward.

MekaTrek is a big idea - there are a lot of features we're still keen to explore and implement. Verification of registered mechanics was a highly discussed option, as well as more sophisticated similarity algorithms to aid predictions and advice to users.

Built With

Share this project: