STA191Data Visualization
Topic 6 · Case Study & Reporting  ›  Week 12

Project Supervision & Data Collection

No new tools this week. Instead: supervision, progress review and group discussion on your case study. This is the week to find problems while there's still time to fix them — two weeks before submission, not two days.

This week you will…

  • Review your group's progress against a clear checkpoint.
  • Validate your questionnaire before collecting more responses.
  • Collect and clean your primary data.
  • Bring draft results to a supervision session and get feedback.
Task pages: 📄 Group project brief & rubric 🎤 Presentation brief & rubric

12.1  What this week is for

Week 12 is set aside for discussions, supervision and progress reviews on your case-study report and presentation — including data collection, questionnaire validation and group discussion. Class time is yours to use; your lecturer is there to unblock you.

● Two weeks left — and a break in between

After Week 12 comes the mid-semester break, then Week 13 (final preparation and the Final Test) and Week 14 (submission and presentation). In practice, this is the last full week of supervised project time. Anything still broken now is your holiday homework.

12.2  Progress checkpoint

Be honest with your group. Tick what's genuinely done — not what's nearly done.

Should be done by nowWhich taskIf not…
Group of 3–4 formed, roles agreedBothDo it today. Week 14 is too late to be looking for members.
Two clear, measurable objectives writtenReportHighest priority. A vague objective sinks the whole report.
Questionnaire built and pilotedReportSection 12.3 below — do this before collecting more.
Primary data collection underwayReportStart now; responses take longer to gather than anyone expects.
Open dataset chosen for the dashboardPresentationPick one this week — DOSM or data.gov.my.
Draft charts built for both tasksBothBring whatever you have to supervision, however rough.
Storyboard sketched (Week 8)PresentationSix boxes on paper. Fifteen minutes' work.

12.3  Validating your questionnaire

Your report needs primary data, which for most groups means a questionnaire. A flawed question produces flawed data, and no amount of clever charting fixes it afterwards.

Check the content first

  1. Map every question to an objective. If a question doesn't serve Objective 1 or Objective 2, delete it — you'll never analyse it.
  2. Check you can chart each answer. Open-ended text is hard to visualise; scales, categories and numbers are easy.
  3. Have someone outside your group read it — a lecturer, or students from another group. This is the “fresh perspective” tactic from Week 8.

Then pilot it

  1. Send it to 5–10 people first, not your whole sample.
  2. Watch where they hesitate, ask you what something means, or leave a question blank.
  3. Fix the confusing items, then release it properly.
Why pilot? Because once 80 people have answered a badly worded question, you can't un-ask it. Ten minutes of piloting saves your Results section.
✗ Questions that cause problems
  • Double-barrelled: “Do you find the library quiet and comfortable?” — which one are they answering?
  • Leading: “How much did you enjoy the excellent facilities?”
  • Vague: “Do you study often?” — what counts as often?
  • Overlapping options: 0–5, 5–10, 10–15 hours. Where does 5 go?
  • Unanswerable: asking for exact figures nobody remembers.
✓ Questions that produce clean data
  • One idea per question.
  • Neutral wording, no adjectives doing persuasion.
  • Specific and time-bound: “In a typical week, how many hours…”
  • Non-overlapping ranges: 0–4, 5–9, 10–14.
  • Consistent scales — don't switch between 4-point and 5-point.
● Session 1 recap

Be honest about where you are. Every question must serve an objective and produce something chartable. Pilot on 5–10 people before releasing.

12.4  Collecting your data

Your Methodology section must describe the reference population, target sample and sampling technique — so decide these deliberately now, rather than reconstructing them afterwards.

DecideAsk yourselves
Reference populationWho is the study really about? All diploma students at Kampus Raub? Residents of one town?
Target sampleHow many responses are you aiming for, and is that realistic in two weeks?
Sampling techniqueHow are respondents chosen — convenience, random, stratified? Name it honestly; convenience sampling is acceptable if you say so.
InstrumentHow many sections and items does the questionnaire have, and what scale?
● Respect your respondents

Tell people what the data is for and that it's for a university assignment. Don't collect names, matric numbers or phone numbers unless you genuinely need them — and if you do, don't put them in your report. Anonymised data is both safer and easier to write up.

12.5  Turning responses into analysable data

If you used Google Forms, responses land in a spreadsheet already. Everything you learned in Week 5 now applies to your own data.

  1. In Google Forms: Responses ▸ Link to Sheets.
  2. Duplicate the sheet before touching anything. Never clean your only copy.
  3. Trim whitespace, standardise text answers, remove duplicate submissions.
  4. Check for impossible values — ages of 200, negative hours.
  5. Note the blanks. Blank is not zero, and treating it as zero will distort your averages.
  6. Build the pivot tables and charts that answer Objective 1 and Objective 2 — nothing else yet.
Two objectives, two sets of visuals. If a chart doesn't answer one of your objectives, it doesn't belong in the Results section. That single rule keeps most reports under the ten-page limit.

12.6  Your supervision session

Come prepared and you'll get useful feedback. Come empty-handed and you'll get general advice you could have read here.

Bring

  • Your two objectives, written out in full.
  • The questionnaire — printed or on screen.
  • However many responses you have so far.
  • Draft charts, even ugly ones.
  • Your dashboard dataset for the presentation.
  • A written list of what you're stuck on.

Expect to be asked

  • “Which objective does this chart answer?”
  • “Why this chart type and not another?”
  • “How were these respondents chosen?”
  • “What does this result actually mean?”
  • “Who is doing which part?”

12.7  Problems worth catching this week

ProblemFix it now
Objectives too broad to chartRewrite them narrower and measurable. Everything downstream depends on this.
Barely any responsesWiden how you distribute it, extend the window, and lower your target to something achievable.
Questions that can't be visualisedConvert open text into categories, or add a scaled question that covers the same ground.
Using a downloaded dataset for the reportThe report needs primary data. Start collecting immediately — see the brief.
One member doing everythingSplit the work in writing today. Teamwork is a scored criterion in the presentation rubric.
No idea what the story isGo back to Week 8 and storyboard it in six boxes.
● Before you leave this week

Agree three things as a group and write them down: who does what, the date your data collection closes, and when you'll meet over the break. Groups that leave Week 12 without these are the ones scrambling in Week 14.

● Session 2 recap

Name your population, sample and sampling technique deliberately. Clean responses the Week 5 way. Bring real material to supervision. Leave with a plan for the break.

Week 13 →