route in Saint Petersburg
distance matrix visualization
route in Helsinki
Do you like travelling as much is we do? If yes, than our inspiration is clear - provide an easy and convinent way to build your holiday plans. Building up your plans can be time consuming and hard, travellers have to do a lot of studing to choose places in big cities. We like to travel and to visit interesting places all over the world. What we want is to help people to travel happilly and without too much preparations, but still get 100% percent of joy and atmosphere.
What it does
Choose a city, a category of interest (museums, religion, etc) and triplanner is ready to roll. It aquires data from different sources (for now that's google maps and wikipedia) and builds a rating of locations. Than they are split into clusters based on distance matrix and the most popular one is chosen. The idea behind it is that the algorithm try to help user to see as much as possible in a shorter amount of time. For example, if user wants to spend one day in a city of choise, the algorithm should lead her to 3-4 most popular places. On the other hand if she wants to spend a week in a city, triplanner can split all locations into 7 clusters so that in each of them all museums are located close to each other. For now we've implemented a very simple prototype, that can build a route for museums for one day.
How we built it
We've got this idea a few years ago. At that time we were travelling in Vietnam, and thougt that it would be nice to have an application like this. No need to read wikipedia for a long time, to study public transport routes - just pure fun.
Challenges we ran into
Main challenge for this application is a good rating system. Our idea for this hackathon was to go with google, and get location ratings from maps API. However it looks like google's ratings are not good enough for us, so in future we are going to use more datasources for it: facebook, twitter, foresquare, instagram.
Accomplishments that we're proud of
Current application is a basic prototype for web, however it already can build simple routes for one day and one city.
What we learned
For the project like this ratings are the main challenge. Using only one source for them is ineffective and sometimes even useless, even if this source is google.
What's next for Triplanner
Our ideas for future development include:
- use more data sources - social networks like facebook, twitter, etc. This way we can build our own rating.
- build route for many days and different categories
- recommend cafes, restoraunts, hotels
- android, ios apps