My name is Vincent, I have been a datascientist by trade for 10 years and the Chief Research / Data Officer of EXPLORE for the last 3 years.
EXPLORE is a french company, based in Nantes, which activity originally mainly involved a business watch about the main events involving french companies, public offices, and territories. Today we are offering our clients either an access to an extranet allowing searching and browsing any pieces of information we gathered and enquired about, or various APIs and services to integrate our data or BI analysis and ML models with our clients or third-party solutions.
As CDO I built a data-centric infrastructure, integrating various technologies (elasticsearch, no-sql databases, GIS, etc.) and launched a few month ago a project around Neo4J.
Our main graph schema involves companies, the buildings they own and/or occupy, economic and capital links between theses entities, main person of interests, and various projects (e.g. construction and real-estate ) & the nature of collaborations between entities involved in theses projects.
I mainly built this graph by myself to discover the neo4j tech and drivers, reading any piece of doc & tutorials I came accross, and conceived an API to query the graph and show parts of it in a linkurious-inspired component of our extranet platform allowing our users to browse parts of the graph around a company.
I'm currently struggling a bit with more challenging queries & usages (e.g. Check different conditions on parts of a path depending of type of relationships and labels of nodes ) and adding more and more new data, new nodes and relationships types to the graph.