A lot of customers hate shopping because they don't even wonder what they want.
What it does
We've built a native app that relies on AI and some quite complex algorithms to recognize colors, match your items with Zalando’s database and get the recommendations for clothes, which would nicely fit their wardrobe.
How we built it
We have a backend written on Python using Django and Google App Engine. The frontend part is an iOS app, written on the brand new Swift 3.0
Challenges we ran into
Fashion as a service (Zalando)
Accomplishments that we're proud of
We have a backend built with clean architecture, which is scalable and can be used even after the hackathon and contains a complex. A swift 3 app is scalable and contains some image recognition algorithms and provides a comfortable interaction with the REST API.
What we learned
Quite a lot about teamwork in extreme situations. Also we've learnt some interesting things from image recognition, like different color distance algorithms. We really enjoyed the process of thinking of an algorithm and development of the working service itself.
What's next for Zalando Wardrobe
We are planning to add some Machine Learning algorithms to make an app's recommendations fit each user's taste better, based on what sets of clothes they like the most. Nevertheless, Zalando promised an API, which will make the mapping of real clothes with ones in their database possible. This will definetely make the process of adding the clothes to the wardrobe easier - user will just scan a bar code. Also it's possible to aggregate different shops inside one app to make the database of available clothes bigger.