Pattern and regularity detection
#Smart operations
Pattern detection is a fundamental branch of data analysis. It mainly consists of the recognition of patterns and regularities in data to understand specific behaviours.
However, it is sometimes more interesting to detect irregularities rather than regularities within data: Identifying an issue inside your supply chain process, detecting frauds operations or exposing suspicious behaviour within a crowd are concrete, high-value use cases.
Exposing desired anomalies
#Smart operations
However, these irregularities can be seen as rare events and are complicated to detect with a high degree of confidence, making it challenging to take adapted and relevant actions to prevent them.
Our methodology is thus designed to detect this outlier behaviour while avoiding the trap of this scarcity phenomenon. We make the best use of available raw data (structured data such as operation logs or even images & video recording) before the processing and modelisation steps to expose the desired anomalies.
Related
Case
Building Orange’s strategy for data monetization
Case
Designing an algorithm to raise consciousness
Case
7 degrees campaign changes winter driving habits