In this talk, we will go on a journey through Zappi’s data history and how we are using Ruby, a graph database, and a bitmap store to build a unique data engine. A journey that starts with the problem of a disconnected data set and serialised data frames, and ends with the solution of an in-memory index. We will explore how we used RedisGraph to model the relationships in our data, connecting semantically equal nodes. Then delve into how a query layer was used to index a bitmap store and, in turn, led to us being able to interrogate our entire dataset orders of magnitude faster than before.