Kakau Kids

Inspiration

Since entertainment media for kids is often just categorized into one genre, it is often hard for children to find new series and movies that they will like. But also for parents, it is difficult to find good content on streaming platforms that is appropriate for kids.

What it does

Kids are given several pictures to choose from. Based on their selection, content recommendations will show up. Parents, who administer their children's profiles, can specify what content they want their children to see in these recommendations.

How we built it

We used NodeJS for Backend and MySQL as our DB. For our responsive web app, we choosed React due to simplification of our contents recommendation. In addition, since the biggest part of our program is depended on the AI (to analyze user preference while considering the parent's regulation), we also used Python to build the recommendation model, based on semi-supervised learning with random forest.

Challenges we ran into

We tried to build a complete web application, that means, we tried to adapt technical skills which are actually on-field. Naturally, those are hard to develop and modularize in one day. But for future extension of our app, we tried to make our project well organized. In addition, although our project goal is 'user' recommendation, we could not access to the user-specific data. Therefore, we should come up with the recommendation model with small labeled data and it took time.

Accomplishments that we're proud of

As we already mentioned, we were able to build a stabilized and well structured web application which is able to be extended in the future. Our features, from very basic functions as Log in to AI analysis, works fine with real world data and they can actually communicate with our users. We tried to minimize the usage of mocked data which make our app strong enough to be published.

What we learned

Although we learned a lot from coding, we learned mostly about managing the project. We underestimated the time limitation, so that we used plenty amount of time to modularize and structuralize our project. In addition, using the real world data was quite different from analyzing idealized data. We learned how to gather the proper data with organized formats and how to process it to use it for training AI.

What's next for Kakau Kids

As we mentioned, our app is well modularized, so it means that there exists big possibility for the future extension. We are looking forward to adapt OpenAPIs (such as the one provided by Netflix) to make our project more interactive. And we are also trying to get permission from "CommonsenseMedia" which provides advanced sets of APIs on this field.

Video Link:

https://drive.google.com/file/d/1eFEf9MPG2zH5MsGXtx2goxiipiFhnNsI/view?usp=sharing