Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Integrating Wearable Data into Process Mining: Event, Case and Activity Enrichment

Vinicius Stein Dani, Xixi Lu, Iris Beerepoot

TL;DR

Traditional process mining often misses dynamic health and behavior signals that influence personal productivity. The paper proposes three enrichment strategies—event-attribute, case-attribute, and new-event generation—demonstrated with calendar data and Apple Health metrics over eight months. It shows how HRV, sleep, and activity relate to daily work patterns and how wearable-derived events can reveal additional process structure. The work highlights practical opportunities and challenges for personal and human-centric process mining, emphasizing measurement validity and privacy considerations to support healthier, more productive routines.

Abstract

In this short paper, we explore the enrichment of event logs with data from wearable devices. We discuss three approaches: (1) treating wearable data as event attributes, linking them directly to individual events, (2) treating wearable data as case attributes, using aggregated day-level scores, and (3) introducing new events derived from wearable data, such as sleep episodes or physical activities. To illustrate these approaches, we use real-world data from one person, matching health data from a smartwatch with events extracted from a digital calendar application. Finally, we discuss the technical and conceptual challenges involved in integrating wearable data into process mining for personal productivity and well-being.

Integrating Wearable Data into Process Mining: Event, Case and Activity Enrichment

TL;DR

Traditional process mining often misses dynamic health and behavior signals that influence personal productivity. The paper proposes three enrichment strategies—event-attribute, case-attribute, and new-event generation—demonstrated with calendar data and Apple Health metrics over eight months. It shows how HRV, sleep, and activity relate to daily work patterns and how wearable-derived events can reveal additional process structure. The work highlights practical opportunities and challenges for personal and human-centric process mining, emphasizing measurement validity and privacy considerations to support healthier, more productive routines.

Abstract

In this short paper, we explore the enrichment of event logs with data from wearable devices. We discuss three approaches: (1) treating wearable data as event attributes, linking them directly to individual events, (2) treating wearable data as case attributes, using aggregated day-level scores, and (3) introducing new events derived from wearable data, such as sleep episodes or physical activities. To illustrate these approaches, we use real-world data from one person, matching health data from a smartwatch with events extracted from a digital calendar application. Finally, we discuss the technical and conceptual challenges involved in integrating wearable data into process mining for personal productivity and well-being.

Paper Structure

This paper contains 3 sections, 5 figures.

Table of Contents

  1. Introduction
  2. Approach and Illustration
  3. Conclusion and Challenges

Figures (5)

  • Figure 1: Depiction of HRV values across internal project meetings (marked with project leader's initial).
  • Figure 2: Depiction of HRV values across three types of recurrent meetings (marked with first letter of meeting).
  • Figure 3: Section of the discovered process model for workdays with a good night's sleep.
  • Figure 4: Median duration projected onto discovered process model for workdays that include walking.
  • Figure 5: Case attribute values for cases shown in Fig. \ref{['fig:walk_and_work']} (dashed red line shows average values across all workdays).