Suggested patterns as deduced from 8,000 trajectories around California University
Pattern from a few trajectories recorded using the Tour Magica PWA.
In today's world a smart phone can guide you a lot better than any human guide for traversing around a place.
What it does
TourMagica is a app, well a PWA , that helps users in exploring new places, by suggesting paths and patterns followed by people that have travelled before. It remembers users' travel paths, analyzes the different trajectories and picks the one best to follow around in a location.
How we built it
We have used the following technology stack -
For the web app - We have a server running in node.js which handles user sessions, collects data from users and routing of the app. MongoDB is used for database purposes.
For computing the best trajectory given a location, we have used python and it's libraries. We have also used available open source codes due to shortage of time.
Challenges we ran into
The machine learning algorithms which we wanted to use were computation extensive and we didn't have the appropriate hardware and time to train the modules for trajectory clustering. Therefore, we had to settle for other trajectory clustering algorithms for now.
Accomplishments that we're proud of
We made a working demo based on the venue of the hackathon which proves the usability of our app.
What we learned
We learned made a Progressive Web App which was new for all of us. We also studied a lot about trajectory clustering algorithms. The live use case of the hackathon area made the problem a lot more real for us.
What's next for TourMagica
We plan to use better trajectory clustering algorithms as that is the backbone of our app. We plan to introduce the factor of 'time spent by users' in finding trajectory which will be an innovation of its own kind.