Reading #1

Poor Form

Read Healy's introductory chapter from Data Visualization for Social Science:

• Look at Data: What Makes Bad Figures Bad

Use the tag “R1” when you post your assessment of the reading and the questions raised.

Within the introductory chapter of Healy's Data Visualization for Social Science, the importance of data visualization is emphasized. Seeing data visually allows for us to see correlation and relationships between data sets that are not visible by looking at just numbers. However, the outcome of a particular graph can vary greatly based upon the choices made to represent the data. The chapter was very informative in terms of giving good and bad examples of design choices such as lines, shapes, and colors. It was very eye opening in showing examples of the same data being dramatically different in terms of slight representation variations. I was introduced to various problems such as aesthetics, bad data, and bad perception. I've learned that aesthetics is an important part to data visualization in order for people to better understand the data and allow them remember the design itself better. Too much unnecessary "junk" takes away from the data itself, but without some sort of originality, the chart becomes ordinary and easily forgotten. Bad data refers to the way that data can be misinterpreted based upon the amount of context shown while bad perception takes into consideration the effects of a stable baseline and aspect ratio on how a graph is read.

Furthermore, the chapter touches upon the issue of intentional misrepresentation. Data visualization is a great way for allowing people to better understand data, but also it can be used with ill intentions to purposefully mislead the audience to believe in false information. I truly wonder whether there is anything considered truly objective within the world of data visualization. So much of it is created by humans and humans are known to have preferences and even biases. Is there a way to prevent being so? If so, how?

It was interesting to see that although data visualization deals with data which is commonly understood as something quite objective, there is no definitive answer in presenting them. There are strategies to not only make data more easily (or difficult depending on intention) interpreted, but also instill a human reaction: most commonly a certain emotion or opinion.

The reading first examines what factors there are in "bad" data representations. The first is aesthetics that deals with simplifying representation. The more unnecessary visual information, the harder to read the data. However it also raises the question of memory and how merely simple graphics may not be memorable for the general public and how humans ultimately rely on visual stimuli and images to retain the data. Secondly, bad data is discussed where not the best kind of data is selected to represent a relationship.

Ultimately, a lot of data visualization is comparing data in terms of relationships: correlation being the main one. In the translation of data to graphical form, the author has lots of visual ways to intensify or neutralize certain relationships. For example, making trend lines steeper by narrowing the gaps of a certain axis. More interestingly, with optical illusions that can make the viewer omit certain information involuntarily.

Personally, this raises significant issues about ethics. We rarely see just a strand of raw data. Data is our primary way to make sense of the world but they are processed by scientists and graphic designers for the general public to comprehend quickly. I am sure most designers are aware of the power they hold in presenting such information. I fear to what extent the world can be brainwashed with malicious data visualization and how the general public can be more educated and aware that no data is ever objective.

What makes bad figures bad

The author talks about three aspects that make bad figures bad: bad taste, bad data, and bad perception.
Bad taste here could be also understood as over design，designers added too many elements that unnecessary for delivering the message, or even fuzz the message. As the author said, given example Figure 1.4 obviously has too many design features such as the background texture, designed font, and the 3-dimensional effects and its shadow, making it hard to read and compare the data. Another example of bad taste is Minard’s visualization of Napoleon’s retreat from Moscow, and the author cited Tufte’s argument that “ Minard’s can be described and admired, but there are no compositional principles on how to create that one wonderful graphic in a million”, which pointed that Minard’s image contains too much-processed information to be used in daily design. While I’m not convinced by the point that only routine and imitable data images are well designed. Different data convey different information. Minard’s image seems to be a unique study of Napoleon’s retreat, which is not daily data. Therefore, in my point of view, it’s not proper to require every visual data to be able to apply to daily use. The author also compared Monstrous Costs’ by Nigel Holmes, and six kinds of summary boxplots by Tufte, explaining the conflict between designers and audiences, saying that over-designed images give audiences a deeper impression than Tufte’s image which delivers the data much more efficiently. In my opinion, delivering the message to audiences is the meaning of the design, only focusing on the “data-to-ink ratio” is the designer’s arrogant choice which ignoring the acceptors.

In Bad data, the author shows “A crisis of faith in democracy” from the New York Times, which we have discussed in class. This image magnified the decline of Democracy by analyzing data in a misleading way. This is the powerful point of data visualization which gives a convincing fact to indicate some forged political view.

The third part, Bad perception, talks about some bad design choices that might confuse audiences, such as 3-D format, junk-free plot, and aspect ratios. Here we see the 3-D format image again, and it really delivers wrong information. Since the column chart is meant to compare just 1-dimensional data, while the 3-D format is creating 3-dimensional volumes, enlarge the original data ratio, therefore it’s an incorrect format.

The reading uses good and bad examples to demonstrate how important it is to consider human perception when designing visual representation of data. I agree that many infographics out there are misleading to the society, so it's crucial for us designers to stay honest with the data information instead of trying to build a strong argument. I'm curious to see if my peers know anything about ggplot and its usefulness because the author mentions it a lot in the reading (it's my first time hearing it).