I actually found this visualization while searching for data for the team project, and it looked like a good candidate for a blog post. Here is where it can be found https://howmuch.net/articles/how-much-must-earn-to-buy-a-home-metro-area
This graph is published on the website is used to show major metropolitan areas across US and how much it takes to buy a house in the certain metropolitan area. This graph is very weak visualization in my opinion because of the following factors:
1. The presentation of the map itself has a nice 3d effect going on. We have discussed this in class and came to conclusion that 3D for any kind of graph is just unnecessary complication and only distract the reader from the message.
2. The scale is nice thing to provide on the legend but it doesn’t match the map. San Francisco for example peaks at 147k, yet the graph of San Francisco is higher than legends 150K. I mean it can be caused by 3D effect, but I did measure it with ruler and it is actually out of scale.
3. Uneven load on the sides compared to the middle of the map. San Diego and Los Angeles as well as Philadelphia and Washington are obstructing each other’s views and make it even harder to read.
4. Bar charts are made in form of the cones and it makes them much harder to read. The pointy ends of cones are not very distinguishable from the map itself. In addition cones tend to distort the proportions.
5. Color combination of the whole thing is just bad, and I am not only talking from aesthetic point of view. The contrast between map itself and bar graphs are not very distinctive, so end of the graphs actually blend in in to the map itself, very bad for visual representation of information. Also it looks like west coast bar graphs look darker then east coast bar graphs, yet according to the legend the color is not a property of bar graphs, colors should be the same across the map, yet they look different. Probably it has something to do with shadow effect of the 3d map.
6. Creators of the visualization probably realized all the visual downfalls to some degree so they actually labeled each bar graph with its value, so my question is “what is the point of visualization if you have to rely on labels to show the values of each data point?”, I don’t think this constitutes a good use of visualization.
7. The most confusing part however is visualization of house median price. It is logical to think that green is good and red is bad. But in this case the colors do not have this meaning; at least I hope they don’t. San Francisco is showed in red while Detroit is shown in green and it can be concluded that Detroit is better than San Francisco to buy a house. But Detroit is a ghost town with high unemployment and crime rate. And even considering the lower housing price it might be actually harder to buy a house there with Detroit’s salary.
8. The final issue is that the 2 parameters reported (the median home price and salaries needed to buy a house) are tied together, unless they do some fancy calculations (which they don’t) or take in consideration any differences in property taxation among different states (which they also don’t). In other words 100k a year salary will be required to buy a house in an area with median house price of 550k no matter where it is located on the map. So what is the point to include them both on the map, especially considering how hard it is to read color coded information (the color gradient very crude and map shadowing affect the colors as well ).
In conclusion it is one of the graphs that is better presented in plain text or graphed as series of bar charts next to each other. Now if users have problems with geography it is best to just include the map separately so users can look up places of interest individually.