Hi Nick, amazing series of blog posts, thanks so much for writing them.
it would be really great, if you could create a graph gist for your model (with a tiny subset of the/some data) so that people can experiment with your queries and give feedback. I'm not really sure that the view-id property that you'd check against is such a good graph solution, if you want to check against a collection of view id's you could do something like (written just out of the top of my head) with extract(v in view | v.view_id) as view_ids where none(n in nodes(path) WHERE n.view_id IN view_ids) but I'd rather check if they are connected with a page with page where none(n in nodes(path) WHERE (n)-[:PAGE]->(page) ) Would love to see how that performs. Also I think your analytics queries might be not that performant yet it in cypher. It would be cool if you meet our London peeps anyway to look into rewriting them as server extension (should be doable within 30-60 minutes) and compare the performance, which should rather be milliseconds than seconds. Cheers, Michael On Fri, Aug 1, 2014 at 12:06 PM, Nick Dingwall <[email protected]> wrote: > I've been experimenting with Neo4J to run event analytics on page view > data, and I've been able to do some pathing analysis that's practically > impossible with SQL. I've written a few blogs on the Snowplow Analytics > website. In the first I define a graph structure that allows for sequencing > of page view events. in the second, I walk through how I got data into > Neo4J, so it's probably not very interesting for people who are experienced > with Neo4J! In the third, I actually perform the analysis. > > I'm very inexperienced with Neo4J so I'd love it if any of you have ideas > for new things to try, or ways to improve the queries I've already written, > particularly in the last post! I'd especially love it if you know of a way > to check a NODES(path) collection for nodes with duplicate properties (in > my case, the URL that the View nodes point to. There's also an edge from > the View node to the Page node, and those pages are unique, so > alternatively we could check in any nodes in NODES(path) have edges to the > same Page node...) > > 1. > http://snowplowanalytics.com/blog/2014/07/28/explorations-in-analyzing-web-event-data-in-graph-databases/ > 2. > http://snowplowanalytics.com/blog/2014/07/30/loading-snowplow-web-event-data-into-graph-databases-for-pathing-analysis/ > 3. > http://snowplowanalytics.com/blog/2014/07/31/using-graph-databases-to-perform-pathing-analysis-initial-experimentation-with-neo4j/ > > Thanks to Michael Hunger for his comments on the first post and for > spotting my error! > > Nick > > -- > You received this message because you are subscribed to the Google Groups > "Neo4j" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > For more options, visit https://groups.google.com/d/optout. > -- You received this message because you are subscribed to the Google Groups "Neo4j" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
