Triangle Counting/Clustering Neo4j

Question

I would like to test Triangle Clustering in my Neo4j graph. Here is a sample:

CREATE(a:Person { name: "a" })-[:FRIENDS]->(b:Person {name : "b"}),
(a)-[:WORKS_AT]->(p:Business {name : "Mcdonalds"}),
(b)-[:WORKS_AT]->(p),
(c:Person { name: "c"})-[:FRIENDS]->(a),
(c:Person { name: "c"})-[:FRIENDS]->(b),
(d:Person { name: "d"})-[:FRIENDS]->(a)
return *

MATCH (c:Person {name: "c"}),(p:Business {name : "Mcdonalds"}), (d:Person { name: "d"}),(b:Person {name : "b"})
CREATE (c)-[:WORKS_AT]->(p),
(e:Person { name: "e"})-[:FRIENDS]->(c),
(d)-[:FRIENDS]->(c),
(d)-[:FRIENDS]->(e),
(f:Person { name: "f"})-[:FRIENDS]->(b),
(g:Person { name: "g"})-[:FRIENDS]->(b),
(i:Person { name: "i"})-[:FRIENDS]->(b),
(h:Person { name: "h"})-[:FRIENDS]->(b),
(j:Person { name: "j"})-[:FRIENDS]->(b),
(k:Person { name: "k"})-[:FRIENDS]->(b)
return *

MATCH (g:Person {name: "g"}),(f:Person {name: "f"}),(c:Person {name: "c"}), (e:Person {name: "e"})
CREATE (g)-[:FRIENDS]->(c),
(f)-[:FRIENDS]->(c),
(g)-[:FRIENDS]->(e)
return *

In my sample graph I would like select nodes a, b, c based on their :works_at relationship with McDonalds, then look at those nodes that have a :friends relationship and use those to conduct a Triangle Count. I've gotten a partial answer with:

CALL algo.triangleCount(
  'MATCH (p:Person)-[]-(:Person)-[:WORKS_AT]-(:Business {name : "Mcdonalds"}) RETURN id(p) as id',
  'MATCH (p1:Person)-[:FRIENDS]->(p2:Person) RETURN id(p1) as source, id(p2) as target',
  {concurrency:4, write:true, writeProperty:'triangle',graph:'cypher', clusteringCoefficientProperty:'coefficient'})
YIELD loadMillis, computeMillis, writeMillis, nodeCount, triangleCount, averageClusteringCoefficient

But I'd like to have something closer to what is listed in the stream example in the documentation with a breakdown of nodeId (in this example node.name), triangles, and coefficient.

I have gotten closer with:

CALL algo.triangleCount.stream(
  'MATCH (p:Person)-[]-(:Person)-[:WORKS_AT]-(:Business {name : "Mcdonalds"}) RETURN id(p) as id',
  'MATCH (p1:Person)-[:FRIENDS]->(p2:Person) RETURN id(p1) as source, id(p2) as target',
  {concurrency:4, write:true, writeProperty:'triangle',graph:'cypher', clusteringCoefficientProperty:'coefficient'})
YIELD nodeId, triangles, coefficient 
MATCH (p:Person) WHERE id(p) = nodeId
RETURN p.id as name, triangles, coefficient  ORDER BY coefficient DESC

Echelon_One · Accepted Answer · 2018-08-09 14:49:16Z

0

CALL algo.triangleCount.stream('match (p:Person)-[*1..2]-(b:Business) return p', '[]', {concurrency:4})
YIELD nodeId, triangles, coefficient
MATCH (p:Person) WHERE id(p) = nodeId
RETURN p.name AS name, triangles, coefficient
ORDER BY triangles

Here's the answer I came up with. The key thing I was missing was understanding the difference between triangleCount and triangleCount.stream. Stream actually analyzes the data while the plain triangleCount merely provides stats on performance, counts, etc.

answered Aug 9, 2018 at 14:49

Echelon_One

276 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

Triangle Counting/Clustering Neo4j

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related