Tufte gives a lot of interesting insight in how to represent multiple dimensions of data on a flat 2-D medium. One of the examples that I found very successful in conveying multiple dimensions of data are the graphs of three air pollutant in southern California. The use of small multiples was very effective because once I understood one design, since the structure was repeated for the other images, I got a sense of familiarity and easily was able to comprehend the other images, which gave even more information on the time of day in which the air pollution is happening. Also, I really like the representation of space and amount of pollutant in this map, as it makes it pretty clear by the heights and locations of the bars where the concentration of pollutant is and how high it is.

One example that Tufte used to describe why chartjunk is generally undesirable was the “Diamonds are a Girl’s Best Friend” graph. I agree that the chart has design beyond the bare minimum, however I still think this chart is effective. I can very easily see the trend of the data and the stylistic choices of the graph and the image of the woman makes the data very memorable and I appreciate the creativity to connect how data looks to real world scenarios. I personally think that Tufte might be overly emphasizing the need for charts and graphs to be high information and show a high density of data. For example, I think it’s actually really difficult to determine trends in the japanese weather chart and that there is almost an information overload. Depending on the use case, I think preparing the data and taking averages or doing other kinds of transformation is wise to do before trying to communicate every single point of data to a viewer.