Filmder

You can have a look at your matches! (Just imagine the thumbnails for now.)
Want to log in?
A whishlist may be written while you're not planning a specific movie session.
Here are all your friends!
A simple profile page.
Connecting can be done easily via a QR code. Not typing-hassle.
A main menu. Not exciting but very useful indeed.
Movies are presented with a thumbnail and a short description.
And another one.

Inspiration

Ever wanted to have a movie night with friends? Finding out what you want to watch is the tricky bit. Everyone has their preferences and everyone has their movies they still want to watch.

What it does

Filmder solves this problem. By giving all the participants a swipe-based interface, it finds a movie-match everyone is satisfied with as quickly as possible. This is done by finding recommendations based on what people already like.

This improves the process because discussions are avoided and everything happens flawlessly and peacefully. You don't have to tell that one guy in your group that it is not a good idea to watch The Lord of the Rings trilogy in the extended version, again.

How we built it

The software consists of three parts.

First off, we have an API handler that manages the several APIs like the public IMDb datasets and the OMDb API to find optimal results. The handler can then respond to requests from the rest of the software.

Secondly, we have a server that handles the user groups and the movies they've already considered.

Thirdly, there is the front-end. The application runs on Android devices and ensures a smooth and pleasant user experience while transferring the data to the back-end components.

Challenges I ran into

While the components themselves are quite operational, a lot still needs to be done to integrate all of them.

Additionally, it is very hard to find publicly available and usable data. To get good results, multiple datasets have to be used and processed. Furthermore, it is basically impossible to use any of them for commercial or non-personal interests. This means that the current implementation is a prototype which, in the current form, cannot be extended to a market-ready solution.

Accomplishments that we are proud of

The back-end is very efficient and handles requests flawlessly. Resources are managed very well so that requests to public APIs can be reduced.

What we learned

Sometimes it's not enough to assume that a program is the sum of its parts. It needs to be more than that. Therefore it is important to focus on the integration of all the components into one polished piece of software.

What's next for Filmder

There is still a lot to do. Once all the components have been connected, the main focal point should be to get a solid infrastructure by publicly deploying the servers. After that improvements have to be made on the recommendation system. This is not going to be an easy task as big amounts of data are required.