Creating short fast Proof of Value models
Bringing Data Science capabilities to the fore in a way that that provides real and measurable value to the organisations we engage with is one of our founding principles, and this piece of work highlighted the benefits of a unified approach.
Our client in this case was a distribution network who presented with a need to improve their processes around predictive maintenance, vegetation management, inventory management and street light analytics.
The client had identified a backlog of scenarios and were seeking to prove the value across these instances. This would ultimately drive future investment in tools and the capability to gain new insights. Initially they had had ‘pockets’ of analytics occurring in the business, but no central capability to prove value across a number of these scenarios.
Our main goal was to bring our utilities domain-based data science and architecture experience to the client to enable them to move quickly through the backlog of scenarios over a short eight-week period. To do this we used an agile fail-fast methodology to quickly identify gaps and issues and move forward to another scenario. Our previous experience also enabled us to position solutions and recommendations into their existing platform/capability from any proven scenarios, enabling them to understand how they could implement these learnings into their existing SAP HANA stack.
We aimed to:
- Prove the value to be gained through application of advanced analytics techniques, including machine learning, on a series of qualified scenarios with the client’s data.
- Prove the value across scenarios for predictive maintenance, vegetation management, inventory management and street light analytics.
- Support the internal innovation team to drive a use case for future data investment through proven value outcomes.
We approached this engagement by working closely with the innovation team and business SME’s to source data and understand the challenges they wanted to address. Our onsite team consisted of a Lead Data Scientist, who was supported by our Solution Architect to provide implementation and future state recommendations. Alongside this, the client provided access to an analyst and business SME resources.
The final deliverables on the project were:
- Executed POV’s across 4 scenarios in a fail-fast approach
- Recommendations provided to ITLT on future value potential in identified areas of vegetation management and street light maintenance, and data quality recommendations to consider proceeding with predictive maintenance
- Provided input to further analytics and innovation business case
- POV report covering each scenario, approach used, findings and recommendations
During this engagement we were able to bring our extensive utilities domain knowledge in combination with our data science and architecture experience, to quickly and effectively identify gaps, issues and measurable value held in each POV without wasting time and resources