The existing fleet data was being matched to external third-party data once a month, but due to data inconsistencies, the required matching of data was a time-consuming, laborious task. This matching process is critical for enabling the state fleet team to perform qualitative checks for “Best Buy” evaluations against the pricing provided by manufacturers.​

The Department sought assistance to establish a Machine Learning (ML) Proof of Concept (PoC) for their fleet management workload.​

DataDivers initially worked with the Department to discover the various facets of the fleet management issues and data. We discovered that matching vehicles between their own dataset and the external one was not straightforward, as some vehicles were described differently across the two datasets.​

Upon investigation, DataDivers recommended the use of data science techniques to “fuzzy” match fleet information using different ML and natural language processing (NLP) techniques. To ensure easy access to the “fuzzy” matched results, we provided the Department with a Power BI user interface as part of the PoC.

The PoC demonstrated the potential of data science and machine learning by:​

• Automating the matching process. This has drastically reduced the manual effort required to reconcile datasets, freeing up the Department’s people for higher-value tasks.​
• Improving the reliability and consistency of data alignment through “fuzzy” matching techniques. This has been critical for informed decision-making.​
• Ensuring more accurate and timely data comparison, so the Department can focus on conducting better value assessments when evaluating manufacturer pricing. This can potentially lead to cost savings or improved negotiation leverage.​
• Democratising access to matched data. The Power BI interface enables non-technical users to leverage the ML outputs without needing technical support.​
• Laying the groundwork for future analytics or AI initiatives. The PoC has shown the potential for integrating smart data solutions across other areas within the Department.