Introduction

Learning Objectives

By the end of this section, you should be able to:

• Explain why sequence analysis reveals patterns that summary statistics miss
• Create and visualize state-sequence objects using TraMineR in R
• Compute pairwise sequence distances with optimal matching
• Compare sequence distributions between user groups using PERMANOVA
• Build event logs and produce process maps, dotted charts, and trace explorations using bupaR
• Apply process mining metrics (activity frequency, throughput time, resource analysis) to HCI interaction data
• Choose an appropriate method for comparing sequential behavioral data

1. Introduction

Event Logs as Behavioral Data

Every time a user interacts with a digital system, a trail of timestamped events is recorded: page views, button clicks, menu selections, error messages, undo actions. These interaction logs are among the richest data sources available to HCI researchers because they capture what users actually do, not just what they report doing.

A common first instinct is to reduce log data to summary statistics — total number of clicks, time on task, error count. These summaries are useful, but they throw away the order in which events happen, and order often matters. Consider two users who both visit five pages and make one error:

• User A: Home -> Search -> Product -> Error -> Help -> Checkout
• User B: Home -> Product -> Checkout -> Error -> Search -> Help

Both have identical summary statistics (5 pages, 1 error), yet their experiences are fundamentally different. User A hit an error early and recovered via the help system before completing a purchase. User B encountered the error after attempting checkout and may have abandoned the task. Sequence analysis lets us see and quantify these differences.

What Sequence Analysis Reveals