Web browsing behavior ofers a rich yet underexplored source of data for process mining (PM) research. However, existing web analytics tools primarily rely on server-side logs or unstructured browsing history, which lack detailed and organized event logging needed for efective process analysis. To address this gap, we introduce Webel, a Chromium-based browser extension tailored for PM researchers. Webel captures structured web browsing events by automatically grouping related activities into coherent process traces while tracking tab relationships and navigation patterns. It supports multiple trace modeling strategies, including parent-child tab relationships, per-tab isolation, and session-wide grouping, providing flexible analytical perspectives. Webel also enables seamless export of logs in the XES format for immediate use in PM tools and ofers optional Google Drive integration for cloud-based log synchronization.
CEUR
ceur-ws.org Screencast video Languages, tools and services used Supported operating environment Chrome Extensions API (Manifest V3) Chromium-based browsers Value Webel 1.0 MIT webels.org/install webels.org github.com/elijahcantu/webel webels.org/demo
PM research has traditionally focused on business processes derived from enterprise systems,
with limited exploration of user-level web browsing behavior as a data source. Although Web
server logs provide information on user interactions with specific websites [
© 2025 Copyright for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
PM analysis. Specifically, they do not organize events into meaningful traces with clear process
boundaries [
This gap between raw browsing data and PM-ready event logs presents a significant barrier
to researchers analyzing web-based user behavior. Current tools lack the ability to create logical
web browsing processes [
Webel addresses these challenges by providing a browser extension that enables the automatic grouping of web browsing events into coherent, semantically meaningful process instances.
Webel is a Chromium-based extension specifically designed for PM applications. Unlike conventional web analytics tools that log events as flat, unstructured sequences, Webel implements multiple trace modeling strategies to capture the semantic structure of user browsing behavior. It preserves relationships between tabs and navigation events and exports the resulting event logs directly in XES format, facilitating seamless integration with PM tools.
Webel supports three trace methods to group web browsing behavior into process instances. Each ofers a diferent analytical perspective: • Parent-Child Tab Traces: This method groups browsing activities based on tab relationships. When a user opens a new tab page, it starts a new trace as the parent. If a link from this tab opens in another tab, the new tab inherits the parent’s trace ID. This way, it captures the user’s main tasks along with any subtasks branching of. • Per-Tab Traces: Each browser tab is treated as a separate trace, regardless of how it was opened. This allows for independent analysis of navigation behavior per tab and is particularly useful when examining parallel exploration patterns. • Session-Based Traces: All activity during a single browser session is grouped into a single trace. This models coarse-grained user behavior across a session, useful for longitudinal task analysis or macro-level patterns.
Each trace method is logged simultaneously, allowing researchers to export and compare all three variants in XES format using the Webel popup or the Google Drive sync. These logs are fully compatible with PM tools such as ProM, Disco, Celonis, and pm4py, enabling immediate analysis without preprocessing.
Webel supports optional Google Drive synchronization. Users can enter a Google Drive folder ID or URL they can access. The logs are automatically synced to that folder using OAuth 2.0 and the Google Drive API. This feature facilitates cross-device analysis and shared research workflows while retaining user control over storage and sharing.
All data remains local unless explicitly exported or synced to Google Drive. The extension popup lets users toggle logging on or of, trigger XES exports, or sync to Google Drive at any time. If logging is disabled, reminder notifications and badge indicators prompt users to re-enable it after they start a new browser session or clear their history.
To evaluate Webel’s capabilities, we conducted a case study focused on how a user accessed course-related resources on the Canvas platform, a widely used learning management system in higher education, at the University of Michigan. In this scenario, the user visited the Canvas page for each enrolled course and then attempted to reach all external resources linked from that course page.
Although this browsing pattern is common among students, it involves a complex sequence of navigation steps that make it well-suited for testing Webel’s trace logic. We recorded the user’s browser sessions during this process and analyzed the resulting logs using Disco 1 after applying basic URL-based regular expressions to label each resource.
Figure 1a–c present process models generated from the same set of browsing sessions, each visualized using a diferent trace modeling strategy. The parent-child trace map captures how external resource visits are organized under each course as distinct sub-tasks, revealing the hierarchical structure of student navigation and how main activities generate subtasks. The per-tab model provides insight into how the user structured their navigation across browser tabs, showing clear separation of activity per course and highlighting parallel exploration behavior. The session trace ofers a comprehensive temporal view of the entire session, reflecting the user’s sequential progress through a course and its associated resources, and capturing habitual return points and overall workflow patterns.
The generated models reveal several important insights. Canvas emerges as the central hub across all perspectives, with substantial self-loop activity (228 returns), indicating frequent returns as a primary “home base.” A dual-hub pattern also appears, with Google Services (148 visits) serving as a secondary integration hub, reflecting the role of third-party tools in educational workflows.
Resource usage shows clear functional distinctions: lecture recordings (30 visits) and Slack (14) are heavily used, while assessment platforms like Gradescope (18) and discussion forums like Piazza (51) see moderate activity. Synchronous tools such as Zoom (4) and ofice hours queues (8–9) have lower but consistent usage.
Taken together, the three trace perspectives provide complementary views: parent-child traces reveal task hierarchies and branching behavior, per-tab traces emphasize parallel exploration, and session traces illustrate sequential workflows and macro-level patterns. This case study (a) Parent-child tab process map
(b) Per-tab process map (c) Per-session process map Figure 1: Comparing navigation flows using three trace modeling strategies highlights Webel’s flexibility in supporting multiple trace perspectives out of the box, allowing researchers to choose the most appropriate level of abstraction for their analysis without requiring manual segmentation or complex preprocessing.
To demonstrate Webel’s applicability to single-tab navigation analysis, we examined a focused scenario within Canvas’s file management system. This study was conducted entirely within a single tab and browser session, so we applied the per-tab trace method. Since Canvas does not support subfolder previews, users must repeatedly enter and exit individual folders to explore the file hierarchy. This behavior results in a navigation pattern that is well-suited for analyzing ineficiencies within isolated browsing contexts.
We recorded browsing events during typical file navigation tasks and applied more
finegrained URL-based regular expressions to label each navigation. The resulting event log was
analyzed using Disco and Cortado [
The tool demonstrates robust support for logging web browsing events, and its XES export functionality has been verified for compatibility with leading PM platforms. Webel focuses exclusively on Chromium-based browsers, and logging activities outside the browser is out of scope for the current design. Future work includes adding support for Safari.
Multiple trace modeling strategies in Webel address the need for flexible process instance identification in web browsing and allow researchers to select the most appropriate level of abstraction for their research questions without manual preprocessing.
During the preparation of this work, the authors used ChatGPT and Claude for grammar and spelling checks. After using these services, the authors reviewed and edited the content as needed and take full responsibility for the publication’s content.