USING TIME CARTOGRAMS TO EXPLORE
TEMPORAL DATA: DO THEY WORK?
Rehmat Ullah, Menno-Jan Kraak, Corné P.J.M. van Elzakker
{rehmat, kraak, elzakker}@itc.nl
OVERVIEW
 Problem/context
 Existing solutions
 Time cartograms
 Usability survey
 Results
 Conclusion
 Future work
March 06, 2013
2
GEOGRAPHICAL DATA
March 06, 2013
3
TEMPORAL DATA
Free Movement
Movement along a fixed network
March 06, 2013
4
PROBLEM
 Difficult to get insight into these data
 Interactive visual analytical representations
need to be designed in order to
 Analyze and synthesize these data
 Produce useful insights about phenomena and
systems associated with these data
March 06, 2013
5
EXISTING SOLUTIONS (1)
 Single map approaches
 space-time cubes
March 06, 2013
6
EXISTING SOLUTIONS (2)
 Series of maps
 Flow maps
 Animations
March 06, 2013
7
ISSUES WITH EXITING SOLUTIONS
 Limited in scope
 Memory overload issues
 e.g. Animations
 Visual clutter problems
 e.g. Space-time cubes and flow maps
 Expensive in terms of visual space
 e.g. Flow maps
March 06, 2013
8
TIME CARTOGRAMS
 Map transformation
 Geographic distance  time distance
 Geography/space is distorted accordingly
March 06, 2013
9
TYPES OF TIME CARTOGRAMS
 Network based
 Some work has been done on network
based time cartograms
 e.g. Bies and van Kreveld, 2012; Ramaer, 2011 ;
Chen, 2011; Kaiser et al., 2010 ; Shimizu and
Inoue, 2009
 Free
 Little or no work has been reported on
time cartograms for free movement
 More exciting and difficult
March 06, 2013
10
EXAMPLE OF NETWORK BASED (1)
March 06, 2013
11
EXAMPLE OF NETWORK BASED (2)
March 06, 2013
12
USABILITY SURVEY
March 06, 2013
13
GEOGRAPHIC vs. SCHEMATIC vs. CARTOGRAM
• Effectiveness
• Efficiency
• Satisfaction
March 06, 2013
14
SURVEY EXECUTION
 8 temporal questions/tasks
 Tasks were mainly realistic and were related to
characteristics like identify, locate, and compare in time
 Randomized algorithm was used to select maps
 Test persons answer first 4 questions using the first map
(selected randomly)
 And next 4 questions using second map (also selected
randomly)
 Complexity of the tasks was even for each type of map
 Finally they give a rank to each map independently
ranging from “Very Poor” to “Very Good”
 Created and maintained using LimeSurvey
March 06, 2013
15
TEMPORAL TASKS (1)
 Locate
 Mr. Smith is in Eindhoven to attend a 3-days
workshop at the Eindhoven University of
Technology. On one of the workshop days he
wants to visit another city that is close to
Eindhoven.
Could you please list all cities to Mr. Smith that
are reachable by train within 60 minutes from
Eindhoven?
March 06, 2013
16
TEMPORAL TASKS (2)
 Compare
 Mr. Malik has only 3 days left in the Netherlands
and wants to visit the Burgers Zoo in Arnhem, the
Railway Museum in Utrecht, or the St. Pietersberg
Caves in Maastricht.
He plans to visit a place which is the nearest to
Eindhoven out of these three.
Could you please tell Mr. Malik the nearest place?
 Arnhem
 Utrecht
 Maastricht
March 06, 2013
17
TEMPORAL TASKS (3)
 Identify
 Which pair of stations is closer (in time) to each other?
 (Eindhoven, Oss)
 (Eindhoven, Roermond)
 (Eindhoven, Breda)
March 06, 2013
18
TEST PERSONS
 35 test persons from 21 different
nationalities
 MSc and PhD students, GIP’s staff
members
 Frequent users of maps
 Make use of the Dutch trains
 Little or no knowledge about the
geography of the Netherlands
March 06, 2013
19
RESULTS
March 06, 2013
20
EFFECTIVENSS
100%
90%
Average Accuracy
80%
70%
60%
50%
Geographic
40%
Schematic
30%
Cartogram
20%
10%
0%
Task 1
[Locate]
Task 2
[Compare]
Task 3
[Locate]
Task 4
[Identify]
Task 5
[Locate]
Task 6
[Compare]
Task 7
[Locate]
Task 8
[Identify]
Temporal Tasks
March 06, 2013
21
EFFICIENCY
Average Execution Time (in seconds)
350
300
250
200
Geographic
150
Schematic
Cartogram
100
50
0
Task 1
[Locate]
Task 2
[Compare]
Task 3
[Locate]
Task 4
[Identify]
Task 5
[Locate]
Task 6
[Compare]
Task 7
[Locate]
Task 8
[Identify]
Temporal Tasks
March 06, 2013
22
SATISFACTION
Cartogram Map
Schematic Map
Geographic Map
Very Good
Good
Average
Poor
Very Poor
23
March 06, 2013
CONCLUSION & FUTURE WORK
 Cartograms
 Effective
 Efficient
 Perform better than the other two maps concerning the
temporal aspects
 Create cartograms for free movement data
 Test their usability
March 06, 2013
24