neo4j mean of property for all friends - graph

Question

Having a graph like:
CREATE (Alice:Person {id:'a', fraud:1})
CREATE (Bob:Person {id:'b', fraud:0})
CREATE (Charlie:Person {id:'c', fraud:0})
CREATE (David:Person {id:'d', fraud:0})
CREATE (Esther:Person {id:'e', fraud:0})
CREATE (Fanny:Person {id:'f', fraud:0})
CREATE (Gabby:Person {id:'g', fraud:0})
CREATE (Fraudster:Person {id:'h', fraud:1})
CREATE
(Alice)-[:CALL]->(Bob),
(Bob)-[:SMS]->(Charlie),
(Charlie)-[:SMS]->(Bob),
(Fanny)-[:SMS]->(Charlie),
(Esther)-[:SMS]->(Fanny),
(Esther)-[:CALL]->(David),
(David)-[:CALL]->(Alice),
(David)-[:SMS]->(Esther),
(Alice)-[:CALL]->(Esther),
(Alice)-[:CALL]->(Fanny),
(Fanny)-[:CALL]->(Fraudster)
When trying to query like:
MATCH (a)-->(b)
WHERE b.fraud = 1
RETURN (count() / ( MATCH (a) -->(b) RETURN count() ) * 100)
I want to compute the fraudulence of a user which (as fraud is only either 0 or 1 is defined as the mean of all connected nodes fraud level:
MATCH ()--(f)
RETURN f.id, f.fraud, COUNT(*), COLLECT(f) AS fs
returns the correct number of friends, but is not able to access these i.e. in the collect statement is only accessing the node itself:
╒══════╤═════════╤══════════════╤══════════╤══════════════════════════════════════════════════════════════════════╕
│"f.id"│"f.fraud"│"avg(f.fraud)"│"COUNT(*)"│"fs" │
╞══════╪═════════╪══════════════╪══════════╪══════════════════════════════════════════════════════════════════════╡
│"h" │1 │1 │1 │[{"fraud":1,"id":"h"}] │
├──────┼─────────┼──────────────┼──────────┼──────────────────────────────────────────────────────────────────────┤
│"f" │0 │0 │4 │[{"fraud":0,"id":"f"},{"fraud":0,"id":"f"},{"fraud":0,"id":"f"},{"frau│
│ │ │ │ │d":0,"id":"f"}] │
....
I.e. naively calculating the average
MATCH ()--(f)
RETURN f.id, avg(f.fraud)
will only consider this single node and not the network. How can I consider the social network of a node instead (up to a defined depth, i.e. here 1) to improve the original answer of neo4j percentage of attribute for social network
edit
MATCH p = ()--()
UNWIND nodes(p) AS f
RETURN f.id, f.fraud, COUNT(*), COLLECT({id: f.id, fraud: f.fraud}) AS fs
will return only duplicates of the original node in the list and not the connected nodes:
│"f.id"│"f.fraud"│"COUNT(*)"│"fs" │
╞══════╪═════════╪══════════╪══════════════════════════════════════════════════════════════════════╡
│"h" │1 │2 │[{"id":"h","fraud":1},{"id":"h","fraud":1}] │
├──────┼─────────┼──────────┼──────────────────────────────────────────────────────────────────────┤
│"f" │0 │8 │[{"id":"f","fraud":0},{"id":"f","fraud":0},{"id":"f","fraud":0},{"id":│
│ │ │ │"f","fraud":0},{"id":"f","fraud":0},{"id":"f","fraud":0},{"id":"f","fr│
│ │ │ │aud":0},{"id":"f","fraud":0}] │
edit 2
MATCH p = (source)--(destination)
RETURN source.id, source.fraud, COUNT(*), COLLECT({id: destination.id, fraud: destination.fraud}) AS neighbors
is already pretty close - but lacking the avg function

Answer 1

MATCH p = (source)-[*..3]-(destination)
RETURN source.id, source.fraud, COUNT(*), avg(destination.fraud), COLLECT({id: destination.id, fraud: destination.fraud}) AS neighbors
includes the fraudulence defined as the average