R1

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.

Kierann Healy’s chapter on ‘badness’ in information graphics walks the reader through the different ways a designer can make a bad or deceiving graph that will not be properly interpreted by the viewer. The main message that can be extracted from this reading is that there are many ways in which one can mess up making a graph. The key to not doing that is by, for the most part, keeping the visuals simple, having correct data, and having the perception of that data conveyed effectively. One interesting point that Healy made was that some charts and graphs that are designed to a greater degree can be better for the viewers, even if it takes them longer to read it. The graph is remembered more easily by the viewer and according to Healy, “makes it more fun to look at.” One thing that I wonder is, does the time that it takes a person to read and understand a graph the most important aspect to a graph? Or does that vary with the subject matter? I think that it depends on the data that the graph is trying to convey. A graph should be aesthetically pleasing and interesting to look at, but if the data does not match the aesthetic, then it should not complicate or get in the way of the information that needs to be conveyed. A graph on a lighter topic matter such as music or pop culture might be able to have more embellishments that might make it harder to read right away rather than a graph on politics or climate change, where the message needs to be understood right away.

This article starts talking about whatever the circumstance is a data visualization to be looked at by someone. It also says Data visualization is not just image making but, it is about who, how, and why it is looked at. Visual perception is most important factor to consider. Data visualization is showing relationship between data and human perceptions. If we make this into simple steps it would be; data, perception, interpretation, and graph construction.

Because the field is about perception and interpretation, miss interpretation can easily happen. However, even with such a big risk of data visualization, It is still the best tool to show people data. Show casing data in visual form is more rememberable for human, who are visually trained than any other senses that we have. Visuals help us to find relationships between things. If graphic are honestly harnessed data visualization can be important tool for communication in human society. Making sure not to lean on perception too much is key to make honest decisions while we are making data into visuals.

Tufte says excellence is needed both in Data and Design. Giving viewers great number of ideas in the shortest time, with list ink, and in smallest space. It is most important that the graphic says the truth.

The reading mainly talks about how tools like position, color, illusion, value, length, angles, area, tilt, depth, curvature, volume, motion, shapes, and scale are all useful. However, these tools are sensitive on the circumstances they are used in. depending on what tools are used where the results can turn out extremely different.

I think this article is to open the thoughts for us and that is why it is very vague about things. word choices like bad, good, honest, taste are used often, which makes the article uncritical. However, important thing that it tells us is that we need to be really careful on our decision making when we design graphics to represent data to be truthful and effective.

In Kieran Healy's introductory chapter of his book Data Visualization for Social Science, he informs the reader on how to critically think about data visualization. He opens a discussion regarding how people perceive graphs, what makes some graphs more apparent and accurate than others, and how to develop good judgment about data visualization. He states that human perception often affects the representation of data. People tend to unconsciously alter graphs even before they start plotting the data; trying to decide which graphing system to utilize, or customize labels and colors affect the correctness of the information. A notably interesting point that he makes is that optical illusions can significantly affect the interpretation of a graph. Simple things such as color or form can considerably affect the perception of data in totally unnoticed ways. Color hue, color luminance, and shapes (for example circles) may trick our mind and make it seem as if the gap difference between 2 variables are smaller than they are, or the opposite, broader than what it is. The real task for accurately representing data is to come up with methods that encode the information properly without harming the study. Taking close attention to our unconscious and being honest with the data is harder than what people might think. Human perception may alter data both when trying to represent information and when reading a graph. It is essential that the focus of a graph is solely to convey the information accurately; embellishments such as color or typography may alter data in negative and unexpected ways.

I think there’s a couple messages central to Healy’s writing. First: a successful piece of information design or data visualization must consider the audience specifically, and human perception more broadly. Healy writes that “An image intended for an audience of experts reading a professional journal may not be readily interpretable by the general public.” As for general human perception, he outlines several ways this trumps a designer’s logical intentions - for example, choosing colors they know are different because they selected entirely unique values, yet recognizing that they may still be similar enough that they’re not easily differentiated by the human eye. He also goes into more formal aspects such as of Gestalt principles, contrast, the components of color, and retinal variables. There was a secondary message towards the end of the piece as well: “problems of honesty and good judgement.” Healy discusses - and acknowledges the merit of - both sides of the debate over what the ethical role is of the designer. Do you include all context to avoid risk of bias or skewed interpretation, or permit modifications that suit your greater point? Ultimately, Healy ultimately seems to come down on the side of the zero baseline, which I would agree with given the caveat that this may not always be the right choice, depending on the information itself and necessary additional context. I am tempted to disagree with his assertion that “chartjunk is not entirely devoid of merit,” but ultimately don’t wholeheartedly. I am personally a fan of a simpler aesthetic across design, but he put forth a convincing argument for the other side. I’d be interested in having a discussion on the three things that can make data design bad: aesthetic, substantive, and perceptual. I don’t disagree with any of them being a significant contributor to a poor design, but I’m curious to know which people see as more/less important. Could you fail at two of the three components, and still have a successful piece of work?

Kierann Healy’s chapter discusses of some factors in data visualization that may mislead the viewers. He introduces a claim that the strong aesthetic judgement isn't the most effective visualization method, but maximizing "data to ink" ratio is another element to consider important. From this viewpoint, simplifying is almost all of what we need to make sure in order to let the charts remain junk-free, and thus effective. At this point however, an opposite opinion arises; some viewers find it easier and more enjoyable to look at the embellished charts. All graphs are meant to be reproducible and effective, but personally I am more convinced to the latter. People of the Internet age are exposed to the overabundant data. In terms of the role of data visualization: to curate meaningful information, readability would be the priority. However, even these well-aligned and executed data may be none of the viewers' interest because we look at lots of them.

Visualized data are sometimes info-biased on purpose. Dramatic value is achieved by manipulating x-axis or omitting zero baseline. It might seem tricky or even worse misleading, but it is remarkable that how data be perceived is variable depending on the adjustment of the graph elements. Here the reading suggests graphic designers, or someone else who make charts, to be honest when dealing with charts and graphs. As mentioned on the reading, there are numerous systems—including tips such as popping up, gestalt rules, or colors—to represent data. These makes graphs more thoughtful for viewers when decoding them. But since "often the main audience for your visualizations is yourself", we should remember that data is so easily be distorted in a way the author want to argue.

Avishi Jain

In Kieran Healy’s introductory chapter of ‘Data Visualization for Social Science’, he attempts to convey to the readers a detailed and critical analysis of data visualization, its importance in delivering information to people as well as the extent of its effectiveness in doing so. The writer repeatedly emphasizes the significance of visually depicting data over the mere examination of numerical summaries or tables and mentions that the former is often more memorable and easily understandable. Organizing numerical values into more easily perceivable graphs and charts such as scatter plots can often be more effective and less redundant than simply running one’s eyes over rows of numerical values arranged in columns of categories.

Healy adds that tools can only help so much in creating effective representations of data and much of the success of this process lies in tasteful yet clear and honest communication of real figures. He also emphasizes the importance of incorporating clarity, precision and efficiency into the process of trying to communicate data in an interesting way using design, statistics and matter of substance.Having a properly chosen format and design to represent data is crucial to creating good graphics. It is also important to use words, numbers and drawings in perfect harmony while keeping in mind the complexity of detail.

While explaining to the readers examples of what may constitute bad representation of data, Healy points out that content-free decoration is one of the greatest obstacles while trying to maximize data-to-ink ratio. While I agree with this to a great extent, I also agree with his later statement that “chartjunk is not entirely devoid of merit”. Ornamental charts are sometimes more aesthetically appealing and easier to remember but this may be true only if the data that is being represented goes hand-in-hand with the aesthetics. On the contrary, sometimes excessive decoration may make a chart more confusing and difficult to decipher. For instance, adding additional dimensions to a chart can disrupt the way the data is read and may lead to misinterpretation of values.

The writer also mentions that it is important to understand what kind of graphs are most appropriate for different kinds of data. What I found particularly intriguing about the author’s analysis of perception and data visualization was the fact that the perception of different retinal variables such as hue and saturation may differ from person to person. For instance, a range of colors differing in value arranged in numerical order may not be perceived as equidistant by everyone. It made me wonder that such differences in perception are likely to exist in several other situations. Thus, would it not be important to keep in mind the audience that the data is being represented to? Having knowledge of the audience and a general understanding of how the data will be perceived by a particular group of people could help design more effective representations that would cater to that population. Another interesting subject of discussion in the article was the honesty of data and this may be lost in the visual communication of data in the form of graphs and charts that mislead audiences. Are there laws against representing data in misleading ways and if yes, to what extent are these valid?

In the introductory chapter of Kieran Healy’s book “Data Visualization for Social Science” the power of data visualization is described. Data visualization can be - and usually is much more successful for conveying a better understanding of data to the human mind. Successful data visualization can rely on many things, including aesthetics, accuracy, and understanding of your audience. The design of Data visualization however, is easily thrown off. There are endless nuances one must consider while designing that are based on human perception. These nuances can range from “bad taste”, to psychological human tendencies. How humans perceive things naturally is a huge factor to consider when designing data, but is endlessly complicated. Furthermore successful data visualization must not mislead the viewer by misrepresenting data. The most vital aspect of data visualization design relies on correctness and also the ease at which it can be perceived. Many classic “Bad Design” rules that apply to typography or layouts also apply to data. It makes sense that a graph that has too many frills, can distract from the meaning. Data visualization is difficult because unlike some types of graphic design, when your personal opinions get involved, you can very easily begin to sway the data into a place where it is hiding part of the story. There are endless ways to create a graph and it’s interesting that some are more correct than others. As the designer, you must make a choice about how the data is best represented in relation to the idea you want to get across. Perception is so important to data design, something as simple as shape choice can completely change how a person interprets a visual.

Kierann Healy’s “Badness” represents things you need to avoid when making a graph. The key point of the reading was that there is a noticeable difference for viewers between graphs, because some of them are obviously easier to read than others. Through the reading, the book encouraged me to look for unconscious choices I make while making graphs and thinking about figures. The author explained three factors which lead to badly designed graph. They are related to aesthetics, substantive evident of data and issues with perception. The aesthetic factor presented in the book is emphasized to be relatively easy to deal with on a graph. By getting rid of unnecessary ‘junks’ from figures and keeping perfect ratio of data-to-ink the graph will be truly effective. Using incorrect or unreliable data is another factor which must be taken into consideration when making a graph. A person making it should fully understand data that he or she is using and also be cautious to not mislead viewers with the ‘halo effect”.  The last factor - issue with perception- was described as how to deal with and combine actual data and the shape of a graph. The book explained that both aspects should be visually pleasant for viewers, so they can easily understand it. Default settings are the simplest forms to understand for viewers, but some other charts tend to remain more challenging to read. I agree with the first and the last factor: making figures junk-free and visualizing data in a more perceptive way. However, I’m not not convinced by the second factor, to not mislead viewers with the halo effect. If a person is only visualizing data, they don’t know if the data is biased. Thus, my question is how to verify by oneself if the data is biased or not?

In the first chapter of Kieran Healy’s book “Data Visualization for Social Science” articulates the different  methods and means for  exploring, understanding, and explaining data. He structures the chapter into eight sections bringing the reader on a journey through the strengths and weaknesses of communicating data into a visually designed figure. He discusses that problems in relation to graphs can be one of three categories: aesthetically tasteless, substantively presented subjectively or perceptually misleading or confusing. A notion I found interesting was that while there are positive aspects to following a simplistic approach to constructing graphics, often what they gain in merit from this "data to ink ratio" they loose in memorability. Healy also details the importance of "how much we are letting the data speak to us, as opposed to arranging it to say what we already think for other reasons."

In his example reporting from the New York Times on the stability of democracies polled across citizens in 6 countries, this idea of substantively sound data comes into play. Simply by assigning the values on the y-value to percentages versus the number scale from 1-10 participants were asked to evaluate can completely shift the readability of the data. An important element to keep in mind is the audience for which you are then going to present this data too.

In a quick glance, or to untrained eyes, elements such as additional dimensions, contrast, and spacing can all influence our perceptual process as a viewer. Not intentionally meant to deceive, these perceptions are something that is not within our control. As the interpreter and creator of data visualizations it is important to keep in mind your audience at all times. Initially having heard of the gestalt principles, I was intrigued in the practical ways in which I saw them implemented effectively. I would be curious as to hear as well as see the ways in which my peers can push and expand these rules to allow for effective visual communication in our poster series project.