Where are the crimes?

The author of this data table and the readers are asking the same question, "Where are the crimes?", but for different reasons.

While the author wanted to convey regional differences in crime growth, as readers, we are not sure which part of the table to look at; every cell is given equal "weight".

Judging from this "profile plot", we can conclude:

• the Mid-West (blue line) experienced a crime spurt that is very much worse than the national average (dots) in all categories except forcible rapes and murder
• the West (red line), in general, had crime increases less severe than the national average
• that said, the regional profiles are relatively similar, showing few meaningful regional differences (compared to other profile plots I've seen)

Reference: "Communities Grapple With Rise in Violence", MSNBC.com
Thanks to Maya for sending in the link.

Comments

[Matias]

A bar chart would be more appropiate to show this data, wouldn't it? This fever chart linked the different types of crime and hierarchized them; a bar chart would allow us to compare them without linking.

[derek]

It took me a minute or two to work out what all the black spots were. I gradually realised the ones on the far right were bullet points for some text, but it was a minute more before I figured out that, alone of all the data series, the dots had a key, and that "last" data point down round -2% for aggravated assault was actually the dot of the key, not a last data point with the word "US Avg" next to it. I don't usually recommend a box framing a key, but here it would help, and perhaps changing the dots so they don't look distractingly like the bullet points in the body of the article would be good as well.

[derek]

Here's my version of the chart. I've changed the profile plot to series of dashes, with the US average as a gray diamond. I've also sorted the crimes in descending order of US average increase.

Looking at my version, I wouldn't have said the West (red) was less severe in general: it's best in murder and robbery, but worst or second worst in all other crimes. The one that stands out to me as less severe in general is Northeast: best or second best in all categories.

[Jon Peltier]

I dislike the use of relative percentage change in this kind of chart. It's meaningless without the context of absolute numbers. Also both versions of the chart (the original and Derek's) take too long to figure out. Derek's removal of lines between categories and sorting of categories are slight improvements, but do little to salvage a poorly conceived chart.

A better chart would have been a "bump" type chart with 2004 values on the left and 2005 values on the right, perhaps in number per 100,000 (or whatever similar normalization is used to present rates of such crimes). The points would be connected by lines (one line per category of crime), the 2005 points would be labeled with the series identifier (including the US Avg, so the legend can be excised) and perhaps the percentage change. The entire label could be used, saving the reader the time to translate the acronyms into region names.

[Jon Peltier]

On second thought, my chart description leaves out at least one factor. Let me amend my design to include five smaller charts, one for each category of crime. Within each chart are five series, one for each region and one for US average.

This morning must be a two-cup-of-coffee morning!

[derek c]

Yes, I kept wondering if the Midwest was really such a hotbed of crime, or if it was just approaching the other regions from a low initial position, a conundrum recently discussed by Kaiser in Rushing to judgement and Saving indices?.

[Kaiser]

Thanks for a very productive discussion. These profile charts are very commonly used and especially when there are many categories, they are somewhat difficult to read.

However, I haven't come across many alternatives that are clearly superior. Derek provides one alternative by removing the lines.

But the lines in this chart serve a different function than on most! Here, they are purely visual aids; they carry no information. Without the lines, it is difficult to trace each region across categories.

The problem with those lines is also that they have no information. In fact, they introduce noise (the slope between crime 1 and crime 2 has no meaning).

I agree that if the individual crime rates were handy, plotting those on a Bumps chart would be a big improvement.

[Kelly O'Day]

I have made an Excel stacked bar chart version.

I have used error bars to make the thin bars. the midwest stands out pretty clearly with this type of chart.

Here's the link:

http://processtrends.com/toc_ch2.gif

Kaiser or Derek, I'm using Firefox. Can you tell me how to add a link to your "post a comment"? I don't seem to have the ability to add a link.

...Kelly

[derek]

Kelly, you need to write the HTML markup code yourself. So

<a href="http://processtrends.com/toc_ch2.gif">the bar chart</a>

becomes

the bar chart