Sophie Fu

One particular takeaway I took from this reading was the fact that in interpreting data trends, it is essential for one to look at both the general dataset alongside the graph visual. If you are able to have the dataset on hand, the chances of a visual bias via graph type or coloration and distribution is less likely. On the other hand, if the bare numbers end up being overwhelmingly large in count, it is helpful to have a graph since outliers are easier to identify amidst the data, or other patterns.

The idea of "less is more" can be readily applied to graphs. If there are too many distracting design elements, the data becomes more of a graphic rather than a display of data. It may be visually appealing, however the crucial point—the data— is often lost in the visual hierarchy.
But it is also noted that more confusing, yet distinct illustrative graphs tend to linger on the minds of viewers more. Due to the graphic possessing more personality, be it colors or the strange monster creature of 'Monstrous Costs' by Nigel Holmes, viewers tend to associate more personal feelings over simple data graphs. In my honest opinion, this becomes a case where the designer must determine if he or she would like to prioritize striking imagery that interests and captivates the viewer, or straight and dry data graphs that helps deliver the information more directly and in more clarity.

The average importance of democracy versus decade of birth graph is another interesting case. While it is important to show the full score of ratings, the small intervals that the y-axis holds creates a feeling of great discrepancy between the generations despite the overall rating to be rather high. You can see a great dip between the 30s and 90s generation in Great Britain, but the actual figures are merely somewhere around 7.3 versus 9.1.

Most of the time, I try to avoid using unordered hues as my palette choice for visuals as I believe that it is far more aesthetic to have a sequential or gradual color set. However, it does make sense when you are graphic several groups of data, as showing a randomized yet equal in value set of colors can help eliminate certain instances of unconscious visual bias.