Data Quality Improvement for Enterprise Asset Management

Engineering and Contracting Services

Maintaining accurate and high quality data records is critical to the success any enterprise asset management strategy.


The number of overhead line equipment (OLE) and high- voltage electrification assets needing to be monitored, distributed across 9,408 kilometres


Reduction in time spent analysing data to accurately assess the condition and quality of assets.


Efficiency savings derived from automating manual surveys, processing data and planning maintenance activities


Our client is an early stage growth company specialising in managing complex civil engineering and electrification projects in the railway industry. They needed to reduce the amount of time spent in monitoring and analysing the condition of over 381,000 assets located across 9,408 kilometres of railway infrastructure. The client needed to overcome a challenge where 14 regional business units created their own methods of recording and storing data related to surveys for these assets.


We worked in partnership with two subject matter experts in project planning for the railway sector and geospatial analysis in the transport sector. Together we discovered that all 14 regional business units collected and stored the information we needed in spreadsheets. We built a data lake to store the raw data files and began analysing the patterns with which the regional business units collected and stored information relating to the 381,000 assets. We discovered that the cause of the inefficiency and the prolonged length of time it took perform a survey of the assets came down to a lack of consistency in data governance.

press to zoom
press to zoom
press to zoom
press to zoom

To automate processes we needed to find hidden and consistent patterns in the datasets. We employed a wide range of data mining techniques to find these patterns and discovered that a lot of the geographical records for the assets were wrong. For example, some assets that were physically in Birmingham were showing up in the Atlantic Ocean. We solved this problem with an innovative yet simple set of clustering and algorithmic techniques based on asset identification numbers.


In less than 5 weeks our team delivered a complete solution that mapped over 381,000 assets and components spread across 9,408 kilometres of railway infrastructure. By improving data quality by 90% we unlocked a 94% reduction in time spent assessing the quality and condition of assets in person throughout the railway network. This helped the client generate over £783,000 per year in operational efficiency savings.

Note: Delivered in collaboration with Pang & Chiu and Open Data Consultants.