ISSS608 2016-17 T1 Assign2 Lee Gwo Mey

ABSTRACT

• The purpose of this assignment is to apply interactive data exploration and analysis techniques to discover patterns from the survey results from University faculty members on the use of Wikipedia as a teaching resources.
• From the data exploration, I found that
  • Respondents with lower acceptance level to adopt the use of Wikipedia are currently non-wiki users and who are either Adjunct, Assistant, Associate or Lecturer.
  • Quality of Wikipedia could be a major concern that affect respondents' decision to use it as a teaching resources.
  • Overall, the external environment is conducive in encouraging the use of Wikipedia except in 3 domains where the use is not so well considered among colleagues. The 3 domains are Arts & Humanities, Health Sciences and Law & Politics.
• The URL link to the web-based interactive data visualization dashboard is found in

== DATA SETS ==

• wiki4HE Data Set (https://archive.ics.uci.edu/ml/datasets/wiki4HE)

INTERACTIVE DATA EXPLORATION AND ANALYSIS TECHNIQUE

• Data visualization software tool - Tableau
• Visual analytics techniques - Divergent bar chart

VISUAL ANALYTICS APPLICATION DESIGN PROCESS

Step 1: Identify a Theme of Interest
Exploratory study on the motivation and barriers behind University faculty members' use of Wikipedia as a teaching resource.

Step 2: Define Questions for Investigation
Motivation factors and barriers that contribute to the use of Wikipedia could be broadly classified under (i) perceptions and attitude, (ii) current skill sets, and (iii) external environment. Initial questions designed to discover these factors are:

• What is the general attitude and acceptance level of University faculty members on the use of Wikipedia?
• Does the quality of Wikipedia affect University faculty members' decision to use Wikipedia?
• Is there a relationship between the University environment and the faculty members' attitude to use Wikipedia?

Step 3: Find the Appropriate Data Attributes

• The wiki4HE data set consists of survey responses from 913 faculty members from 2 Spanish Universities (UOC and UPF).
• The survey comprises 43 Likert-scaled questions to assess respondents' agreement to each question.
• The following survey questions and their responses are selected to review if these can provide insights to the investigation questions in step 2 above.

S/N	Qns Ref	Category	Qns
1	BI2	Behavioural Intention	In future, I will use Wikipedia in my teaching activity
2	EXP1	Experience	I consult Wikipedia for issues related to my field of expertise
3	JR1	Job Relevance	My university promotes the use of open collaborative environments in the Internet
4	PEU1	Perceived Ease of Use	Wikipedia is user-friendly
5	PU3	Perceived Usefulness	Wikipedia is useful for teaching
6	QU1	Quality	Articles in Wikipedia are reliable
7	IM2	Social Image	In academia, sharing open educational resources is appreciated

CONSTRUCTING THE VISUALIZATION

Data Preparation & Transformation

• Import raw data file (wiki4HE.csv) into MS Excel
• Create a unique "Respondent_ID" for each survey respondent
• Split the data file into 3 separate MS Excel worksheet tab
  • Demographic - convert the response code (e.g. gender "0") to text (e.g. "Male")
  • Question - include for each question ID, the question category (e.g. "perceived usefulness") and a brief question text (e.g. "wiki is useful")
  • Response - include responses in both numeric format (e.g. "3") and text format (e.g. "neutral")
• Null values - replaced the "?" (questions not answered) with blank fields
• Liker Scale - label each scale as follows:
  • 1-Strongly Disagree
  • 2-Disagree
  • 3-Neutral
  • 4-Agree
  • 5-Strongly Agree
• Reshape data - the current survey response format is in a one-row-per-respondent and one-column-per-question table. I used the Tableau add-in for MS Excel to reshape the format into a one-column format for all responses.
• Although I am focusing my data discovery on the 10 survey questions above. I will still import all 43 survey questions into Tableau. This will allow me the flexibility to interact with the data set and refine my questions during the exploratory phase.

Divergent Bar Chart²

• The divergent bar chart is my chosen visualization technique as it best present the negative, neutral and positive sentiments in each bar relative to each question.
• 6 calculated fields are created in order to create the divergent bar chart.
  • Count Negative

IF[Numeric Response]<3 THEN 1
ELSE IF [Numeric Response]=3 THEN 0.5
ELSE 0 END

INTERMEDIATE VISUALIZATION

FINAL VISUALIZATION

References: