Prototype

Neubau after Dark

What if…

we could easily create recommendations for community-driven events in our new Headquarter in Berlin?

Problem

We asked ourselves if this tool is actually necessary because we could also use Meetup or Facebook for coordinating events but we thought about spontaneous gatherings like meeting for ice cream on the roof terrace which is nothing you would coordinate via Meetup. Would it not be nice to receive personalized recommendations based on types of events you previously attended and liked?

Solution

We created a slim platform on which the users can create events by describing what they want to do and the app extracts automatically time and activity. Of course it can be manually adjusted. The more events created, the smarter the interface becomes because it learns more and more locations and activities based on the user’s input. While browsing through the list of available activities, the user decides to like or dislike the events so that we can afterwards send proactive notifications based on the user’s profile.

Challenges

We tried to shape this prototype as realistic as possible so that we implemented social logins (via Google, Github and more) and a small classifier. We discontinued the classifier approach and decided to fall back to a content-based filtering mechanism.