Anime ratings, what we can tell from them and our prediction

Inspiration

We wanted to dive deeper into this topic because lots of people have trouble choosing what anime to watch next or if a new anime is worth watching. So we wanted to help make suggestions based on anime features. Knowing the answers could help anime writers or producers know what is making certain animes more popular and have higher ratings that please viewers. This helps them adapt and allows them to know what is trending or what features might make a better anime.

What it does

Produce graphs that tell us about the anime dataset and predicts anime ratings.

How we built it

• Read in the dataset and clean up the data
• Merge required dataset and one-encode required columns
• Defines methods for our plots
• Create a decision tree regressor machine learning model to predict ratings
• Split data 80/20 to train and test
• Determine mean squared error
• Create visualizations and graphs to interpret our data findings
• Observe results and edit code as necessary to improve visualizations

Challenges we ran into

Dataset was too large so we had to create and work with samples.

Accomplishments that we're proud of

Created many graphs and an ML that is able to predict the rating of an anime based on its feature.

What we learned

• Gender does not play a role in how much people rate their anime, and in what rating scores they give.
• We are able to predict the ratings of an anime based on its features ## What's next for 'Anime ratings, what we can tell from them, and our prediction'
• Do more research as to what other variables may affect an anime’s rating
• Perform better statistical analysis in determining important variables
• Create an anime rating system that allows for users to input new anime and compute their ratings to determine whether a new anime is worth viewing or not.

Built With

• python