Wednesday, 14 October 2015

Stockholm meetup got a bit out of hand

Tonight we had a lovely meetup in lovely Stockholm at our friends of HiQ. They have a wonderful office and event area overlooking downtown Stockholm, a perfect setting to sit back and talk graphs. So we did. My friend and colleague David Montag did a wonderful talk, and I tried to tell people how to really NOT mess up their Neo4j project. The presentation is over here:

But somewhere along the way, I also started talking about my lack of a social life (so sad!!!) and the fact that I have a bit of fun with graph karaoke. I know - it's stupid. But so about half an hour before my talk, someone says "Can you do that for "Roxanne"?" ... and I am like - being a big Police fan and all - alright, Challenge Accepted.

So here's what I did:
  • I grabbed the lyrics from AZLyrics
  • I put it into this google sheet.
  • I made the google sheet available to the public (important).
  • I downloaded that sheet as a CSV file, and grabbed the URL from my downloads folder.
  • and then ran this cypher query over it:
 load csv with headers from "" as csv  
 with csv.Sequence as seq, csv.Songsentence as row  
 unwind row as text  
 with seq, reduce(t=tolower(text), delim in [",",".","!","?",'"',":",";","'","-"] | replace(t,delim,"")) as normalized  
 with seq, [w in split(normalized," ") | trim(w)] as words  
 unwind range(0,size(words)-2) as idx  
 MERGE (w1:Word {name:words[idx], seq:toInt(seq)})  
 MERGE (w2:Word {name:words[idx+1], seq:toInt(seq)})  
 MERGE (w1)-[r:NEXT {seq:toInt(seq)}]->(w2)  

And that's it. Graph is ready. Next I find the song in Spotify, play it a few times to rehearse, and then we get this:

So there you go. I hope you enjoyed it as much as I did. And don't forget to register for GraphConnect next week!!!



No comments:

Post a Comment