AniRec: Anime Recommendation System

Homepage
Anime Recommendations for Naruto

Inspiration

Fans of anime have always struggled to find something new to watch. While it is plausible to ask others for suggestions, it is not time efficient and everyone has a taste of their own. This is a problem. We wanted to create an easy to use web app that allows each and everyone to find their personal recommendations tailored to their liking.

What it does

AniRec gathers information of top 8 matching anime based on the similarity of user search. The web app then displays those anime and their user rating on the webpage in a simple, easy to read format.

How we built it

We used Flask combined with Html and Css to make an interactive web app.

The app used AnilistPython module to make API call in order to retrieve anime data from a website called AniList. Based on the anime user input, we used a content-based recommendation algorithm and ran a similarity score test based on genres to search for anime similar to the one user inputted. We also used a .csv file to temporarily act as our "database".

Challenges we ran into

We had several challenges throughout this project of creating our own web application. The main one being finding a topic which we are capable of building within 24 hours but would also have real world impact and benefits.

We have also faced and overcame technical challenges. Initially, we didn't have sufficient skills and proficiencies to turn our idea into an MVP. Therefore, in the last 24 hours, we have engaged in a continuous loop of constant Googling, learning, and coding.

Accomplishments that we're proud of

We are proud of being able to come up and implement an idea in 24 hours, and perform a satisfactory video demonstration that introduced our creation.

What we learned

We learnt the importance of the practicality of an idea. While we had many other good ideas at the start of the hackathon, most of them were not feasible for us to build given the short time limit. We have also obtained various technical knowledge such as using Flask to make a web app, API, and creating a rating algorithm. Most importantly, we learnt how to work effectively as team.

What's next for Anime Recommendation System

We hope to implement a comment section under each anime entry to encourage community interactions, create user accounts to facilitate individualised recommendations, add code that allows users to report recommendations so the algorithm can be finetuned, and overall, improve the scoring algorithm.

To further optimise our query time, we are planning to cache past queries so that future users will have a smoother experience.

Built With

anillist
css3
flask
html5
javascript
python
sci-kit

Submitted to

SYNCS Hack 2022

Created by

I worked on the back-end of the project. I looked into content-based recommendation algorithms and applied it to our project. Scikit and numpy helped a lot, as they were instrumental in implementing the similarity scoring.

Jacob Lee
I assisted on the backend, it was my first time using an API so there was quite a learning curve. My main job, was actually recording and editing/constructing the video demonstration, which I found quite fun.

David Vuong
I worked on the algorithm to recommend user similar anime , this include but not limited to retrieving and filtering data from API, transform existing data into usable form for our similarity calculations.

Huan Vo
I worked on the front end using HTML and Flask to connect to the back end

dang nguyen
I worked on front-end and back-end of the project. I used HTML and CSS to design the website UI and used Flask web framework to retrieve user input and feed it to our algorithm, then render the list of recommended anime on the website.

Mu-Wei Chung
I worked on the presentation and helped with the devpost

Dylan Liu