We attended Neo technology’s Neo4J graph database event GraphConnect in London a few days ago. As we mentioned in a previous post, we’ve been using graph technology from the beginning, as we’re convinced the problem we’re solving can’t be efficiently solved by traditional relational databases, which basically work like spreadsheets. Try to describe your whole supply chain in Excel (or whatever your favorite spreadsheet software is), from the grower to the finish product, and you’ll soon be either:
- creating one tab per supplier, but how many tabs should you create in total? and what if a supplier shows up at multiple tiers?
- or tracking all the connections on a single tab, and wind up with macros and pivot tables unable to tell you anything
Graph databases allow us to store and query supply chains as they are: a large intertwined network of suppliers.
It was really exciting for us to see that Supply Chains are not the only topic that is a good fit for graph databases: Alicia Power (@apowers411) from Fino Consulting and Mar Cabra (@cabralens) from ICIJ had great stories to tell at GraphConnect!
Alicia Powers is a data scientist who used Neo4J to map and analyze people food behavior, and use that to make healthy and realistic recommendations. Diets are usually about a one-size-fits-all approach: her speech title – Who cares what Beyonce ate for lunch? – actually refers to Beyonce-endorsed 22 days vegan diet. The issue with that kind of diet, is that people usually don’t really like what they have to eat, which leads them to quickly dropping off.
Powers’ approach is actually to identify, for each individual, people that have similar tastes but a healthier behavior, and use that to recommend changes to that person’s eating habit: if you spend your day snacking on candies and crisps, there’s little chance you’re suddenly going to switch to quinoa and organic pineapple. But if you get recommended to sit down and have a real lunch (even if you pick burger and soda), you’re more likely to accept, and you’ll certainly reduce snacking, thus improving your health!
Mar Cabra, head of ICIJ’s Data and Research Unit, was one of the most expected speakers. She explained how the ICIJ Used Neo4j to Unravel the Panama Papers. Although the ICIJ had previously exposed leaks such as the Swiss Leaks or Luxemburg Leaks: they had received from the anonymous whistle-blower 650 times more data than they had in any other leak! They could have been through the documents one by one, relying on their intuition and post-it notes, but they actually recognized the data for what it was: a densely tangled web of people and companies.
Once you have the data in a graph instead of a long list of documents, you don’t need much effort to spot that the same person appears on several company documents, or that the same pattern connects celebrities with offshore companies. The database is actually now available online so that anybody can browse it, search it or download it.
As one of Forbes contributors recently pointed out, some information « just can’t be neatly organized into uniform columns and rows » which is where graphs make it possible to connect the dots between offshore companies, choose the diet that suits you over Beyonce’s, or map your entire supply chain.
