Data Structure Visualization

1
Data Structure Visualization
Vikrant Colaso Kunal Garach Anuj Shah
2
Why use animation for visualizing Data Structures?

• Instead of students mentally running the
  algorithm, the animation helps rapid perception
  of the data structures.
• They can also dynamically portray the changes
  in data structures as the algorithm evolves.
• User can control speed of execution and thus
  learn at his own pace.

3
Previous Work
Numerous attempts have been made to prove that
Visualization helps in learning.
Our Study
Testing Retention of Information Gained using
Different Media. (Animation, Standard Text and
Both)
4
Tool Used
5
Tests

• Phase 1
• Students from CS 2604 were given a time limit
  of 20 minutes to learn both DFS and BFS using
  either the text, tool or both.
• A 10 minute, multiple choice test was taken at
  the end to evaluate their understanding of
  procedural as well as conceptual aspects of the
  algorithms.
• Phase 2
• After a period of 15 days a second test was
  conducted to test the retention.
• Tests were designed using HTML and JavaScript in
  order to record the time taken on each question.

6
Data Collected

• Demographic Survey
• Name, GPA, Prior knowledge of Graph and Graph
  searching techniques.
• Post Test Questionnaire
• Understanding of algorithms, prediction of score,
  helpfulness of presented
• material, comments on the material.
• Tests
• Time taken on each question, total scored on each
  category of questions.

7
Data Analysis

• Using Spotfire and Excel.
• ANOVA, T- tests.

8
Results
82 confident that there is a significant
difference between TT and Text
Average scores for the Text Group is the least.
9
Decrease in Average Scores
Results Contd.
10
This indicates a trend that degradation for the
Text group is the most.
Results Contd.
79.83 confident that there is a difference in
average scores for the second test
11
Results Contd.
Better
61.50 confident that there is a difference in
average procedural scores for the second test
This indicates a trend that degradation for plain
text is the most.
12
Results Contd.
99.88 confident that the Text Tool combination
is rated as the most helpful.
13
Results Contd.
88.62 confident that the average GPA on
TextTool was less than that on Text
14
Is this relevant?
91.68 confidence that the average time taken on
conceptual questions is less for tool.
15
Summary of Trends in data

• A poor average score for those who used the
  text on the 2nd test.
• A poorer average score on procedural questions
  as well for those using the text.
• The textual material was the least appreciated.
• GPAs of students on the text were on an average
  higher than those using both
• the text and the tool.

16
Conclusions

• General Comments
• Almost all the students liked the Tool and felt
  that it was a novel method for teaching.
• Those who got the Text felt they may have done
  better with a graphical visualization of the data
  structures.
• Remembering the algorithms was difficult for
  almost all the students however those using the
  tool could remember some of the steps.

17
Further Work
Recommend that students be given a longer
time/Assignments with the tool to notice a
trend. Recommend testing with review of material
before taking the second test. Recommend using a
tool with speed control, history and other
enhancements. Studies should be conducted with
a larger number of students and questions.