We present the results of a study that identifies common activity patterns through analysis of eye-tracking data and the event logs of the popular authoring tool, Protégé. Informed by the activity patterns discovered, we propose design guidelines for bulk editing, efficient reasoning and increased situational awareness. Methodological implications go beyond the remit of knowledge artefacts.
1 of 20
Downloaded 13 times
More Related Content
Constructing Conceptual Knowledge Artefacts: Activity Patterns in the Ontology Authoring Process
1. Constructing Conceptual Knowledge Artefacts:
Activity Patterns in the Ontology Authoring Process
Markel Vigo, Caroline Jay, Robert Stevens
University of Manchester (UK)
CHI 2015, Seoul (Korea)
@markelvigo
markel.vigo@manchester.ac.uk
13. Findings: interaction log data
• Interaction events account for 65% of events
while authoring events are 30%
• The top 3 events (entity selection, description
selection and invocation of editing menu) account
for 56% of events
6
12
19
23
28
39
47
61
82
113
139
142
182
199
267
314
332
426
617
960
1004
1405
2793
Back
Undo
Get explanation
Entity renamed
Set property
Entity dragged
Property addition
Entity deleted
Load ontology
Hierarchy collapsed(i)
Save
Description selected(i)
Run reasoner
Hierarchy collapsed
Convert into defined
Hierarchy expanded(i)
Class addition
Entity selected(i)
Hierarchy expanded
Entity edited:finish
Entity edited:start
Description selected
Entity selected
0 1000 2000
14. Findings: eye-tracking data
The class hierarchy is the pivotal area
• Index of the ontology
• External memory
Transitions between AOIs
from
to
Ann−Usage
Class hierarchy
Description
Explanation
File menu
Pop up
Edit Entity
Prop. hierarchy
Ann−U
sageC
lass
hierarchy
D
escription
Explanation
File
m
enu
Pop
up
EditEntityProp.hierarchy
0
1000
2000
3000
4000
5000
6000
15. Findings: eye-tracking data
The class hierarchy receives users’ attention
45% of the time
0
100
200
300
400
File
m
enu
Ann.−U
sage
C
lass
hierarchy
D
escription
Popup
Editentity
Prop.hierarchy
Explanation
AOI
time(sec)
17. Implications: from raw data to workflows
Workflows can be automatically identified
raw data
cleaning
merging
filtering
workflow
detection
~7K rows
~200 rows
• Different authoring styles
• Time distribution per workflow
• Identification of confounding variables
18. Implications for design
• Support for bulk editing
• Anticipation of reasoner invocation
• Automatic detection of authoring problems
• Make changes to the inferred hierarchy
explicit
19. tl;dr
• Identification of activity patterns when dealing
with complex interactive artefacts
• Interaction log data + eye gaze data
• Data-driven
• Application on knowledge artefacts
20. Markel Vigo, Caroline Jay, Robert Stevens
University of Manchester (UK)
@markelvigo
markel.vigo@manchester.ac.uk