Unbored

Inspiration

Uuuuuuuh, how can we say it. yes. The project is based on the challenge proposed by Nexthink.

What it does

Unbore us. I learn what you like and recommend you activities that you would like. I am not very intelligent for the moment but it can change, at least you're not alone, I'm with you :).

How we built it

We searched for different datasets related to differents kind of activities that humans do and can do. Then we analyzed and filtered the data we wanted in the app before pushing them into a MySQL server using python to parse and format the data into a coherent form. Then, we made a little web app that asks you some questions about what activities you like doing and store them. Then it recommends you different activities based on what you liked . You can re complete the questionnaire to have better further recommendations. The recommendations are based on the K-NN distance algorithm.

Challenges we ran into

Dealing with datasets where the datas are spreads between different files. Working from home with your colleague is complicated, it takes more time to solve a simple problem.

Accomplishments that we are proud of

The web app doesn't crash.

What we learned

We've learned how valuable working in person with colleagues is.

What's next for Unbored

Maybe we can use different technologies to store and compute the data. We wanted to propose a highly distribued and fault tolerant system to handle all the machine learning process and storing by using technologies like AKKA, Apache Kafka, Cassandra DB, etc. All the ML process could be handled into actors that get the datas from a Cluster Kafka. And in general, MORE ML for the project.