Recommend the best movies for a viewer

This model recommends the best movies for a viewer based on the how the viewer, and other viewers, rated movies. It is not based on any demographic information about viewers or categorizing viewers by other means.

The dataset was provided by GroupLens Research . The data consists of hundreds of thousands of movies rated (from 1 to 5) by thousands of users, over a period of time.
Submitted By
Snap Analytx Inc.
Date Submitted
Source Code
Java (Mahout)
Collaboartive Filtering
Analytic Engine
Download Users Preferences Table
Download List of Movies Table
Data Dictionary
Attribute Name Brief Description Type Range Low Range High
User IdThe unqiue id of each userInteger1943
Movie IDThe unqiue id of each movieInteger11682
RatingRating assigned to a movie by an userInteger15
Time Stamp Time when the user rated the movie expressed as unix seconds since 1/1/1970 UTC. (not used for generating recommendations). Integer 874724710 893286638
  • The dataset was provided by GroupLens Research Checkout the site for additional details about the data
  • The ratings were collected over various periods of time between 1997 and 1999
  • Each of the 943 users has rated at least 20 movies
Users Preferences Table
User Id Movie Id Movie Name Rating
  • There is one row for every movie rated by a user; for each user there are as many rows as the number of movie he/she rated.
  • The table shows the movie preferences for five customers only. The full set has 100,000 ratings by 943 users for 1682 movies.
  • The "Movie Name" shown here is not part of the data dictionary shown on the privious tab; it is a "lookup" of the movie id from the Movie Database.
  • The data file (click the "Download Table" button) used by the model does not have column headers.
Movie Database
Movie ID Title Release Date Genre Imdb Link
Imdb Link
  • The "Movie ID" here is the same as the one in the "User Ratings Data" tab.