based on Nintendo game data, cluster by features and get similar games

recommendation.py

import pandas as pd
from sklearn.cluster import KMeans

# Read the video games dataset into a DataFrame
data = pd.read_json('../data/nintendo-games.json')

features = data[['meta_score', 'esrb_rating', 'genres']]

# convert categorical features to numerical using one-hot encoding
features = pd.get_dummies(features, columns=['meta_score', 'esrb_rating', 'genres'])

# use k-means clustering to group similar games together
km = KMeans(n_clusters=20, n_init=10)
km.fit(features)

# add the cluster labels to the dataframe
data['cluster'] = km.labels_


def recommend_games(favorite_games, num_recommendations=5):
    """
    Recommends games based on input favorite games using KMeans clustering.

    Parameters:
        favorite_games (list): List of favorite game names.
        features (list): List of feature names to use for clustering.
        num_clusters (int): Number of clusters to use in KMeans clustering. Default is 5.
        num_recommendations (int): Number of recommended games to return for each favorite game. Default is 10.

    Returns:
        dict: A set containing a list of names of recommended games from the same cluster as the game.
    """

    # Identify the clusters of the favorite games
    favorite_games_data = data[data['title'].isin(favorite_games)]
    if favorite_games_data.empty:
        raise ValueError("No data found for the favorite games.")

    # Find other games in the same clusters as favorite games
    recommended_games_set = set()
    for favorite_game_row in favorite_games_data.itertuples():
        similar_games = data[data['cluster'] == favorite_game_row.cluster]
        similar_games = similar_games[similar_games['title'] != favorite_game_row.title] # Exclude favorite game        
        similar_games = data.loc[similar_games.index]  # Get the full rows of similar games
        similar_games_sorted = similar_games.sort_values(['meta_score', 'msrp'], ascending=False)
        recommended_games = similar_games_sorted.head(num_recommendations)['title'].tolist()

        recommended_games_set.update(recommended_games)

    # Remove the favorite games from the recommended games set
    recommended_games_set.difference_update(set(favorite_games))

    return recommended_games_set


# Example usage:
favorite_games = [

]
recommended_games = recommend_games(favorite_games)
print(recommended_games)