Skip to main content

Strengthen your risk management

Leverage datascience and advanced statistical analysis to strengthen your risk management and anomaly detection.

Case study

Use advanced statistical analysis to find connections not listed for a Utilities actor


Network performance is a major issue for all resource distribution management companies. Indeed, losses on the network represent a loss of income for the company, and consequently for the taxpayer.

Some physical connections are not found in information systems. These are wild connections or connections that are poorly informed. These connections are not billable or maintainable. As a result, they can represent a significant loss.

Our team therefore initiated a prototype over a period of 2 months to see if it was possible to detect these unlisted connections.



In order to detect these connections, we have targeted geographical areas of underconsumption, modelled using machine learning methodologies. We enriched the approach with Open Data to precisely target suspect buildings. Finally, in order to facilitate the investigation of the functional experts, we have created a visualization interface restoring the suspicious connections, with a 3D vision of the surrounding context. 


Key success factor 

  • Open Data Valuation
  • Regression models for modelling area consumption
  • Mass geocoding of branch addresses
  • Visualization of results on Google Maps: satellite and 3D vision to better explore the results of the approach


Promising approach highlighting consumption anomalies, verified in the field. It should be noted that a strong sensitivity to geographical coordinates - and consequently to the quality of the available addresses - was a major source of difficulty. 

Launch of a project to improve the reliability of connection addresses


Cybersecurity, Fraud detection, Outliers detection

Our teams of experts in Data Science and Cybersecurity support you in strengthening your risk management: risks of cyber-attacks, fraud or risks linked to the presence of aberrant data in the information system. Detecting such risks as early as possible is vital and often within reach, thanks in particular to the analysis of weak signals present in your data. Our technical expertise is also complemented by a strong and multi-sector business knowledge that allows us to respond to the specific problems of each of our customers.


Heka is the ecosystem of Artificial Intelligence solutions developed by Sia Partners. These advanced Data Science solutions come from years of development experience and support of our customers. Our developed industrial tools and insights allow Sia Partners to address recurring business issues and support value creation across multiple sectors.