Fixing Bad Visualizations
Presentation for Data Viz SG / HHSG meetup
Topics to be covered
- Introductions
- Data viz quiz
- Q1: Why visualizations.
- Q2: Chart types. Data aggregation / reduction.
- Q3: Exploratory vs Presentation. Perception.
- Q4: Validation. Data scale.
- Q5: Data scales. Channels.
Hi!
I'm Chi-Loong. Nice to meet you!
I am part of DVS and help run the Data Viz SG and HHSG meetup groups.
I run my own visualization studio at V/R doing work for clients.
I also work as product / engineering head at Mogul.sg
I also teach data visualization as an adjunct lecturer at SUTD / SIT.
Data Viz Quiz
These quiz questions that I'm going through were taken from the last SIT CSC3007 Information Visualization Sem 3 final quiz.
It'll be a fun exercise to go through these and discuss!
Please note that as data viz is both an art + science, the answer can be subjective and dependent on context!
Q1. Why is it not enough to just simply rely on summary statistics to understand datasets?
Why should we still do visualizations, even if they are rough exploratory ones?
Q2. The following chart shows the salary range for a profession.
- Is this a good visualization idiom for this dataset? If yes, why? If not, why not, and how would you improve it?
- The visualization only shows the data from one profession. How would you visualize a more complicated dataset that has 20 professions?
- What if you want to show even more information – the salary range for 20 professions across 2 years?
Assume that the new visualization will not be interactive. Beyond that, feel free to make any other assumptions as necessary.
Q3.You overhear this conversation about two friends talking about this New York Times Covid visualization.
"I hate this Covid visualization. It's so hard to read and understand!" Says Alan.
"It's not designed for accurate reading of the data," said Betty.
- Do a critique of the visualization and explain in your own words what you understand from Alan's and Betty's perspectives.
- If you were asked to make a more accurate and easily read visualization of Covid19 cases using the same dataset, how would you redesign this visualization? Assume that the new visualization is for print and will not be interactive. Beyond that, feel free to make any other assumptions as necessary.
- Explain how and why your suggested visualization is more accurate than the New York Times one.
Q4.The Singapore Ministry of Health unveiled a geospatial interactive visualization of Covid cases on thier Covid-19 report portal on October 1st 2021. A screenshot of the visualization is below.
The map classifies areas into 5 color bands. The darkest color scale indicates that 230 to 560 covid-19 cases have visited the area in the past three days. A screenshot of the color scale is below:
Do a critique of the visualization and explain in your own words whether the visualization is good or bad.
How would you improve the visualization, if any?
Q5.Take a look at the following chart that shows relationships between different currencies, their coupon rate, and the year when their coupon matures. This version is not interctive.
Do a critique of the visualization and explain in your own words whether the visualization is good or bad. Feel free to justify your critiques with your assumptions.
How would you improve the visualization, if any? You can make the final chart interactive.