Skip to content

Data Card: MovieLens Rating Explorer

Dataset Source

This project uses the MovieLens Latest Small Dataset from GroupLens Research.

The dataset contains movie information and user rating activity. It is used here for educational data analysis.

Purpose

The purpose of this project is to identify highly rated movies while considering how many user ratings support each result.

A movie with a high average rating from only a few users may be less reliable than a similarly rated movie supported by many users. The interactive app lets the user choose a minimum rating-count threshold before viewing results.

Records and Grain

The project uses two related tables:

  • movies.csv: one row represents one movie.
  • ratings.csv: one row represents one user's rating of one movie.

The tables are connected with the movieId column.

Variables Used

movies.csv

Variable Description
movieId Unique identifier for each movie
title Movie title and release year
genres One or more genres assigned to the movie

ratings.csv

Variable Description
userId Identifier for the user who submitted a rating
movieId Identifier connecting the rating to a movie
rating User's rating of the movie
timestamp Time the rating was recorded

Processing and Analysis

The Marimo app loads both CSV files into an in-memory SQLite database.

SQL joins the movies and ratings tables using movieId. The query then:

  1. counts the ratings for each movie;
  2. calculates each movie's average rating;
  3. filters movies using the selected minimum rating count;
  4. sorts the results by average rating;
  5. returns the top ten qualifying movies.

Results

With a minimum of 50 ratings, The Shawshank Redemption (1994) had the highest average rating: 4.43 out of 5 from 317 ratings.

Increasing the threshold to 100 ratings removed movies with fewer than 100 ratings from the ranking and replaced them with movies supported by more user feedback.

Limitations

  • The dataset represents MovieLens user activity, not all movie viewers.
  • User IDs do not provide demographic information about the people who rated the movies.
  • Average ratings are opinions from MovieLens users, not objective measures of movie quality.
  • A higher minimum rating count provides more supporting evidence, but it may exclude highly rated movies with fewer ratings.