orchard/sonymusic neo4j hackathon

Cypher Load CSV

USING PERIODIC COMMIT 10000
LOAD CSV WITH HEADERS FROM 'file:///Tracks.csv' as line
WITH line,
	CASE   
        WHEN toFloat(line.TEMPO) < 20 THEN "Larghissimo"
        WHEN toFloat(line.TEMPO) >= 20 AND toFloat(line.TEMPO) < 40 THEN "Grave"
        WHEN toFloat(line.TEMPO) >= 40 AND toFloat(line.TEMPO) < 60 THEN "Largo"
        WHEN toFloat(line.TEMPO) >= 60 AND toFloat(line.TEMPO) < 80 THEN "Adagio"
        WHEN toFloat(line.TEMPO) >= 80 AND toFloat(line.TEMPO) < 120 THEN "Moderato"
        WHEN toFloat(line.TEMPO) >= 120 AND toFloat(line.TEMPO) < 180 THEN "Allegro"
        ELSE "Presto" 
    END as NamedTempo,
    CASE 
        WHEN toFLoat(line.DANCEABILITY) < .25 THEN "S"
        WHEN toFLoat(line.DANCEABILITY)  >=  .25 AND toFLoat(line.DANCEABILITY) < .5 THEN "M"
        WHEN toFLoat(line.DANCEABILITY)  >= .5 AND  toFLoat(line.DANCEABILITY) < .75 THEN "L"
		ELSE "XL" 
    END as Dance,
    CASE 
        WHEN toFLoat(line.ENERGY) < .25 THEN "S"
        WHEN toFLoat(line.ENERGY)  >=  .25 AND toFLoat(line.ENERGY) < .5 THEN "M"
        WHEN toFLoat(line.ENERGY)  >= .5 AND  toFLoat(line.ENERGY) < .75 THEN "L"
		ELSE "XL" 
    END as Energy,
    CASE 
        WHEN toFLoat(line.VALENCE) < .25 THEN "S"
        WHEN toFLoat(line.VALENCE)  >=  .25 AND toFLoat(line.VALENCE) < .5 THEN "M"
        WHEN toFLoat(line.VALENCE)  >= .5 AND  toFLoat(line.VALENCE) < .75 THEN "L"
		ELSE "XL" 
    END as Valence,
    CASE 
        WHEN toFLoat(line.INSTRUMENTALNESS) < .25 THEN "S"
        WHEN toFLoat(line.INSTRUMENTALNESS)  >=  .25 AND toFLoat(line.INSTRUMENTALNESS) < .5 THEN "M"
        WHEN toFLoat(line.INSTRUMENTALNESS)  >= .5 AND  toFLoat(line.INSTRUMENTALNESS) < .75 THEN "L"
		ELSE "XL" 
    END as Instrumentalness

MERGE (tmp:Tempo {tempo:NamedTempo})
MERGE (d:DanceAbility {score:Dance})
MERGE (e:Energy {score:Energy})
MERGE (v:Valence {score:Valence})
MERGE (i:Instrumentalness {score:Instrumentalness})
MERGE (t:Track {trackId: line.TRACKID})
SET 
	t.trackName=line.NAME,
	t.isrc=line.ISRC,
    t.popularity=line.POPULARITY

MERGE (t)-[ht:HASTEMPO]->(tmp) SET ht.value = toFloat(line.TEMPO)
MERGE (t)-[dt:DANCEABLE]->(d) SET dt.value = toFloat(line.DANCEABILITY)
MERGE (t)-[et:ENERGETIC]->(e) SET et.value = toFloat(line.ENERGY)
MERGE (t)-[hat:HAPPINESS]->(v) SET hat.value = toFloat(line.VALENCE)
MERGE (t)-[it:INTRUMENTAL]->(i) SET it.value = toFloat(line.INSTRUMENTALNESS)

WITH t, line
MATCH (a:Artist {artistId:line.ARTISTID})
MERGE (a)-[:PERFORMED]->(t)

orchard-sme-hackathon-neo4j-cohort-notes.md

https://ec2-18-221-218-113.us-east-2.compute.amazonaws.com:7473/browser/

$ ssh -i alex_sony.pem [email protected]

Query

MATCH (t1:Track)-[v1:HASTEMPO|HAPPINESS|ENERGETIC|INTRUMENTAL|DANCEABLE]->(value1) WHERE id(t1) = 5026
WITH t1, v1
ORDER BY value1.order
WITH t1, collect(v1.value) as values1
MATCH (t2:Track)-[v2:HASTEMPO|HAPPINESS|ENERGETIC|INTRUMENTAL|DANCEABLE]->(value2) 
WHERE id(t2) = 18
WITH t1, values1, t2, v2
ORDER BY value2.order
WITH t1, t2, collect(v2.value) as values2, values1
WHERE size(values2) = 5

RETURN t1.trackName AS from,
       t2.trackName AS to,
       values1,
       values2,
       algo.similarity.cosine(values1, values2) AS similarity

sample_cohort_track.csv