TrackSpotter

• 

  Kanban in Frankfurt

• 

  Home Screen

• 

  Play Screen

• 

  Hacking on the Train

• 

  Hacking in DB Frankfurt

• 

  More Hacking on the Train

• 

  Team Selfie

• 

  Hacking in Paris 4:30am...

• 

  Pitching to the Judges

Inspiration

We've all have been on a train, staring out of the window and looking for a fun way to pass the time. It turns out we can use our time to help Network Rail get more accurate data on the position of their line-side assets (such as signage, portacabins, and more) using a game community, originally inspired by the reCAPTCHA system to identify parts of images.

What it does

Our two game interfaces help to filter large amounts of GPS-taged video into stills from various sources (driver cab, drone and potentially smartphone cameras), down to a manageable data sets that specify the location of line-side assets, thus reducing time taken looking at irrelevant stills, and splitting the effort between the gaming community. Our solution provides the exact data requested by network rail, including GPS location (±3m), asset type and accuracy score.

How we built it

We built it using Python, MongoDB, React and Javascript.

Challenges we ran into

Activation: High volume engagement is key to success. The more images that are viewed the more assets are identified with a higher degree of confidence. Therefore, the game has to be fun and rewarding for users.

Accuracy: Ensuring accuracy of position data up to <3meter, while it is possible to visually spot assets up to 20-30 meters away.

Flexibility: We aim for a flexible solution that can be applied to different data sources and identifiable assets on the railway.

Accomplishments that we're proud of

Awards for Entrepeneur & 3rd best hack at the EU train. Fantastic team spirit and cooperation throughout the weekend & a working demo!

What we learned

That a simple, spontaneous idea can turn into a useful and viable solution if you have the right resources and mindset.

What's next for TrackSpotter

We are looking forward to great pitch and are happy to present at the HackTrain conference on 22 November 2018.

Built With

• javascript
• mongodb
• python
• react