Magic Recommendations

Static view of graphs page.

Inspiration

Machine learning is very popular nowadays, and it was quite a challenge to take this with so little experience. We love to come out of our comport zone and learn something new !

What it does

The program implementation should recommend products based on a user activity and best-selling recently viewed products.

How we built it

The implementation of the application was done in python. Using Jupyter notebook with libraries like pandas, plotly, numpy

Accomplishments that we're proud of

We achieved good results on the scoreboard, despite programming experiences of teammates. Also, we did an exciting analysis of dataset visualized in HTML page.

What we learned

We learned that it is better to think first and then do something (xD) and team leader should think twice before choosing team in the future.

Conclusion and Future Work

Overall we are happy with the result, at first we tried machine learning with no success because of the lack of experience, and then we moved to easy algorithms, like manual filtering through data. Our next achievement is a working HTML page with embedded graphs of analyzed dataset. For example, the best hours when the user is active, the best hours for each category and many more. For the future work there is still so many improvements in method how we can approach the task. Mainly usage of deep learning method is a good start. But most important is to think about the dataset and how to effectively prepare it for usage.

Happy Coding Team. Sincerely, team leader Peter, a.k.a. Monnte.