Location-Aware Path Alignment in Process Mining

Location-aware log data is an untapped source of information that promises new business analysis insights. This is in particular the case for business processes that can be linked to sensor data such as RFID or WiFi signals. Technically, this question can be formulated as a special type of alignment problem, which is well known in process mining. In this paper, we formalize the alignment problem for spatio-temporal event data. Our contribution is a novel algorithm that finds sensor IDs that travel together on the basis of their location information. Questions centered around spatio-temporal event logs may include all kinds of movements, such as customers in shops highlighting 'Hot and Cold areas' or tracking of material and goods in a production plant. For this paper, we choose a specific challenge for retail companies, which is to find out if customers are alone or visit the shop together with family or friends. Therefore, the algorithm is tested using positioning-data of a retail shop from the fashion industry. Our results highlight the benefits of location-based process mining by showing its applicability in real scenarios.