Sensor Data Analytics PoC
Proof of concept for analysing very large sensor datasets with Qlik ODAG and Vertica, built as a two-person delivery team and selected over a competing architecture.
Scope
The client had a global fleet of machines reporting technical and financial event data into a large database. They wanted to explore whether Qlik could support efficient analysis across very large volumes of detailed sensor and transaction records.
The proof of concept compared a Qlik ODAG and Vertica architecture against an alternative implementation from another vendor.
What I did
I built the Qlik analysis layer: overview applications for broad analysis and selection narrowing, connected to detail applications generated dynamically through Qlik ODAG. The detail applications queried lower-level records from Vertica only after users had narrowed the context enough to make the query practical.
I also worked on Vertica optimisation, including projections needed to support the summary views used by the overview apps.
Outcome
The Qlik ODAG and Vertica approach was chosen by the client because it handled the required data volumes and provided the analytical workflow they needed.