[Neo4J][Cypher]A basic Movie Recommendation Engine in Cypher

[Florent Empis]

Hi,

I have just released the code for a MovieLens based recommendation engine.
I got the inspiration from http://markorodriguez.com/2011/09/22/a-graph-based-movie-recommender-engine/ who did pretty much the same in Gremlin.

The barrier to entry in Gremlin was too high for me, so I settled for Cypher.

I did everything in Cypher: loading data into the graph, then querying the graph to return the recommendations.

There's also a class dedicated to create full-text indexing on movie title, which mixes Cypher and pure neo4J commands.

The code is here https://github.com/FlorentE/neo4j-experiments

And a blog/tutorial post is here http://blog.bidul.es/post/2012/06/21/Neo4J/Cypher-Recommendation-engine4

Any feedback and advices are most welcome!

BR,

Florent

[Michael Hunger]

Sweet, well written post and good explanations.

two comments,

# you could forgo the manual indexing if you set up auto-indexing on the title and name and configure the auto-indexer to use a fulltext-index.

# could you add line-breaks to your cypher statements, as your blog page is very narrow it is hard to read them w/o scrolling

Would be cool to have a neo4j-console (console.neo4j.org) loaded with (parts) of the dataset to play around with the recommendation queries.

Thanks a lot

Michael