Perceptual Biases in Data Visualization
There are several websites discussing factors that may cause people to perceive information with bias. One of the common factors is that people’s perception of the magnitude of objects (e.g. circles) is dependent on surroundings and could be disproportionate to the actual area. Therefore, if numbers are represented by those shapes, people will perceive the quantities inaccurately. Additionally, color choices also affect the perception of a graph. For example, different brightness might be associated with different weights of the data. Some colors could be associated with different meanings in different cultures.
Graphics aren’t to be seen, but to be read and scrutinized.
Data Visualization and Digital Humanities.
As Lin mentioned in her lecture, there are two purposes of visualization: presentation and exploration. Those two functions exist on a continuum for any single graph, and both are valuable for digital humanities. Usually, it starts with exploration: researchers use visualization to help them find trends in data and conclusions. Then, the researchers present their points to the audience through data visualization. Since graphs are often more informative than summary statistics but more concise than lines of words, it is a common choice in digital projects to visualize humanistic data.
However, as mentioned in the previous section, there are many pitfalls of visualization that could cause the audiences to misperceive the data than the researchers originally intended. So researchers must avoid such practices in their graphs.
Additionally, many choices have to be made when producing a graph. Should a visulization emphasize decoration or functionality? Do we want a dense graph with more information or a light one with minimal information? Is it better to choose charts that are hmore familiar to people (e.g. bar charts) or some more original data visualizations to catch attention? My suggestion is to always consider the context and purpose of the graph so that it fits into the general picture of a digital project. Be sure to include sufficient information (e.g. what are the variables) along with any graph to make it more informative to readers.
2 thoughts on “Data Visualization: How Not to Lie with Pictures”
Comments are closed.
I like how you mentioned that considering the context around your visual is really important, like who is going to be looking at it and how much they will know about the subject. If you want to a reach a broader audience then it would be wise to keep the graph simple and easy to understand without any prior knowledge on the subject.
I like how you discuss the unconscious biases that impact viewed perceptions when viewing data visualization graphics. It seems like it is always important to keep in mind who your audience is and how they will unconsciously interpret data that is being presented. I think it is also interesting to think about how visualization changes based on whether you are exploring or presenting.