Created
September 9, 2023 01:37
-
-
Save rjurney/8bdd16741d7ec5864eb09386c03ee0c3 to your computer and use it in GitHub Desktop.
Feature engineering using JanusGraph for Relato's business graph - customer recommender system in 2014
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
// Use this to start up a session | |
conf = new BaseConfiguration() | |
conf.setProperty("storage.directory", "/Users/rjurney/Software/marketing/titan/data") | |
conf.setProperty("storage.backend", "berkeleyje") | |
graph = TitanFactory.open(conf) | |
// Get a graph traverser | |
g = graph.traversal() | |
// Various centralities to use as features - JSONize and save | |
partnershipCentrality = g.V().out('partnership').groupCount().by('domain').next() | |
partnerPrettyJson = new JsonBuilder(partnershipCentrality).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/partner_centrality.json").write(partnerPrettyJson) | |
doublePartnershipCentrality = g.V().out('partnership').out('partnership').groupCount().by('domain').next() | |
doublePartnerPrettyJson = new JsonBuilder(doublePartnershipCentrality).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/partner_partner_centrality.json").write(doublePartnerPrettyJson) | |
partnerCompetitorCentrality = g.V().out('partnership').out('competitor').groupCount().by('domain').next() | |
partnerCompetitorPrettyJson = new JsonBuilder(partnerCompetitorCentrality).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/partner_competitor_centrality.json").write(partnerCompetitorPrettyJson) | |
competitorCentrality = g.V().out('competitor').groupCount().by('domain').next() | |
competitorPrettyJson = new JsonBuilder(competitorCentrality).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/competitor_centrality.json").write(competitorPrettyJson) | |
competitorCustomerCentrality = g.V().out('competitor').out('customer').groupCount().by('domain').next() | |
competitorCustomerPrettyJson = new JsonBuilder(competitorCustomerCentrality).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/competitor_customer_centrality.json").write(competitorCustomerPrettyJson) | |
competitorPartnerCentrality = g.V().out('partnership').out('competitor').groupCount().by('domain').next() | |
competitorPartnerPrettyJson = new JsonBuilder(competitorPartnerCentrality).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/competitor_partner_centrality.json").write(competitorPartnerPrettyJson) | |
customerCompetitorCentrality = g.V().out('customer').out('competitor').groupCount().by('domain').next() | |
customerCompetitorPrettyJson = new JsonBuilder(customerCompetitorCentrality).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/customer_competitor_centrality.json").write(customerCompetitorPrettyJson) | |
// New simple degree stuff I realized I hadn't added sooner, although some may be equivalent | |
degree = g.V().group().by('domain').by(both().count()).next() | |
degreeJson = new JsonBuilder(degree).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/degree.json").write(degreeJson) | |
inDegree = g.V().group().by('domain').by(inE().count()).next() | |
inDegreeJson = new JsonBuilder(inDegree).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/in_degree.json").write(inDegreeJson) | |
outDegree = g.V().group().by('domain').by(out().count()).next() | |
outDegreeJson = new JsonBuilder(outDegree).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/out_degree.json").write(outDegreeJson) | |
outDegreePartnership = g.V().group().by('domain').by(out('partnership').count()).next() | |
outDegreePartnershipJson = new JsonBuilder(outDegreePartnership).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/out_degree_partnership.json").write(outDegreePartnershipJson) | |
inDegreePartnership = g.V().group().by('domain').by(inE('partnership').count()).next() | |
inDegreePartnershipJson = new JsonBuilder(inDegreePartnership).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/in_degree_partnership.json").write(inDegreePartnershipJson) | |
outDegreeCustomer = g.V().group().by('domain').by(out('customer').count()).next() | |
outDegreeCustomerJson = new JsonBuilder(outDegreeCustomer).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/out_degree_customer.json").write(outDegreeCustomerJson) | |
inDegreeCustomer = g.V().group().by('domain').by(inE('customer').count()).next() | |
inDegreeCustomerJson = new JsonBuilder(inDegreeCustomer).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/in_degree_customer.json").write(inDegreeCustomerJson) | |
inDegreeInvestment = g.V().group().by('domain').by(inE('investment').count()).next() | |
inDegreeInvestmentJson = new JsonBuilder(inDegreeInvestment).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/in_degree_investment.json").write(inDegreeInvestmentJson) | |
outDegreeInvestment = g.V().group().by('domain').by(out('investment').count()).next() | |
outDegreeInvestmentJson = new JsonBuilder(outDegreeInvestment).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/out_degree_investment.json").write(outDegreeInvestmentJson) | |
competitorDegree = g.V().group().by('domain').by(both('competitor').count()).next() | |
competitorDegreeJson = new JsonBuilder(competitorDegree).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/degree_competitor.json").write(competitorDegreeJson) | |
// end global centralities | |
// Particular to a customer | |
customerConnectedness = g.V().has('domain','oreilly.com').as('x').out('customer').in('partnership').where(neq('x')).groupCount().by('domain').next() | |
customerConnectednessJson = new JsonBuilder(customerConnectedness).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/oreilly_customer_connectedness.json").write(customerConnectednessJson) | |
customerConnectedness = g.V().has('domain','gainsight.com').as('x').out('customer').in('partnership').where(neq('x')).groupCount().by('domain').next() | |
customerConnectednessJson = new JsonBuilder(customerConnectedness).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/gainsight_customer_connectedness.json").write(customerConnectednessJson) | |
customerConnectedness = g.V().has('domain','relato.io').as('x').out('competitor').out('customer').both('partnership').where(neq('x')).groupCount().by('domain').next() | |
customerConnectednessJson = new JsonBuilder(customerConnectedness).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/relato_customer_connectedness.json").write(customerConnectednessJson) | |
customerConnectedness2ndBi = g.V().has('domain','oreilly.com').as('x').out('customer').in('partnership').both('partnership').where(neq('x')).groupCount().by('domain').next() | |
customerConnectedness2ndBiJson = new JsonBuilder(customerConnectedness2ndBi).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/oreilly_customer_connectedness_2nd_bi.json").write(customerConnectedness2ndBiJson) | |
customerConnectedness2ndBi = g.V().has('domain','gainsight.com').as('x').out('customer').in('partnership').both('partnership').where(neq('x')).groupCount().by('domain').next() | |
customerConnectedness2ndBiJson = new JsonBuilder(customerConnectedness2ndBi).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/gainsight_customer_connectedness_2nd_bi.json").write(customerConnectedness2ndBiJson) | |
customerConnectedness2ndBi = g.V().has('domain','relato.io').as('x').out('competitor').out('customer').both('partnership').both('partnership').where(neq('x')).groupCount().by('domain').next() | |
customerConnectedness2ndBiJson = new JsonBuilder(customerConnectedness2ndBi).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/relato_customer_connectedness_2nd_bi.json").write(customerConnectedness2ndBiJson) | |
customerConnectedness = g.V().has('domain','totango.com').as('x').out('customer').both('partnership').where(neq('x')).groupCount().by('domain').next() | |
customerConnectednessJson = new JsonBuilder(customerConnectedness).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/totango.com_customer_connectedness.json").write(customerConnectednessJson) | |
customerConnectedness2ndBi = g.V().has('domain','totango.com').as('x').out('customer').both('partnership').both('partnership').where(neq('x')).groupCount().by('domain').next() | |
customerConnectedness2ndBiJson = new JsonBuilder(customerConnectedness2ndBi).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/totango.com_customer_connectedness_2nd_bi.json").write(customerConnectedness2ndBiJson) | |
competitorCustomerConnectedness2ndBi = g.V().has('domain','totango.com').as('x').out('competitor').out('customer').both('partnership').both('partnership').where(neq('x')).groupCount().by('domain').next() | |
competitorCustomerConnectedness2ndBiJson = new JsonBuilder(competitorCustomerConnectedness2ndBi).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/totango.com_competitor_customer_connectedness_2nd_bi.json").write(competitorCustomerConnectedness2ndBiJson) | |
customerConnectedness = g.V().has('domain','welink.com').as('x').out('customer').both('partnership').where(neq('x')).groupCount().by('domain').next() | |
customerConnectednessJson = new JsonBuilder(customerConnectedness).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/welink.com_customer_connectedness.json").write(customerConnectednessJson) | |
customerConnectedness2ndBi = g.V().has('domain','welink.com').as('x').out('customer').both('partnership').both('partnership').where(neq('x')).groupCount().by('domain').next() | |
customerConnectedness2ndBiJson = new JsonBuilder(customerConnectedness2ndBi).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/welink.com_customer_connectedness_2nd_bi.json").write(customerConnectedness2ndBiJson) | |
competitorCustomerConnectedness2ndBi = g.V().has('domain','welink.com').as('x').out('competitor').out('customer').both('partnership').both('partnership').where(neq('x')).groupCount().by('domain').next() | |
competitorCustomerConnectedness2ndBiJson = new JsonBuilder(competitorCustomerConnectedness2ndBi).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/welink.com_competitor_customer_connectedness_2nd_bi.json").write(competitorCustomerConnectedness2ndBiJson) | |
// Skip me when redoing things until the end | |
partnershipLoopCentrality = g.V().repeat(out('partnership').groupCount('m').by('domain')).times(3).cap('m').next() | |
partnershipLoopCentralityJson = new JsonBuilder(partnershipLoopCentrality).toString() | |
new File("/Users/rjurney/Software/marketing/titan/data/partner_loop_centrality.json").write(partnershipLoopCentralityJson) | |
// Closeness Centrality | |
// closenessCentrality = g.V().as("a").repeat(both('partnership').simplePath()).emit().as("b").dedup().by(select("a","b")).path().group().by(limit(local, 1)).by(count(local).map {1/it.get()}.sum()) | |
// subGraph = g.E().hasLabel('partnership').subgraph('subGraph').cap('subGraph').next() |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment