Visiting a new town is daunting. You don't know where to eat where to meet people at the bar so we often ask 'Where should I go?'. As a team mostly coming from outside of Waterloo we didn't know where to eat in Toronto. Of course existing solutions like Yelp give you a lot of information about the places you have to choose from, but I'm busy and don't have time to look through usually what you really want to know
and existing solutions do a good job of giving you options, however, what we usually ask for whe
What it does
How I built it
We used a large host of APIs including Twitter, Instagram, Uber, Google Maps, Indico, and Yelp. We used Firebase for our realtime backend and linode servers for data processing. Aside from that, the whole application was built with Python + some front end love :)
Challenges I ran into
We ran into a couple of issues with data query encoding and Jinja when linking flask with the frontend. The most substantial issue was confusion over latitude & longitude, but we luckily sorted that out.
Accomplishments that I'm proud of
Our queries using Yelp and Indico are very accurate theme wise, and return results that don't just emulate the "top 3" that Yelp or Google Maps would return. The implication of our social media analysis was clearly evident!
What I learned
We learned how to use several APIs, implementations to topic modeling, and how to use Linode!
What's next for Urban Adventure
Expand to more data sources, make the search backend granular, reinforce the topic modeling backend, and revamp the design of the project. We stayed up almost all weekend, but there's still ton we wish we could have done!