Tuesday, January 31, 2012

learning from good visuals

I find I spend a lot of time here discussing less than stellar information visualizations and how they can be improved. But there is also much to be learned about data viz best practices by examining good visuals and understanding why they are effective. I paused on one such graphic last week (see original article):


Here's a quick rundown of what I like about the above visual:
  • Everything is labeled (titles, axes, important values, data source), so there's no question about what you're looking at. The overall title explains what the visuals are intended to convince you of (so there's no guessing!). The points they want to highlight (2011) are shown clearly and draw attention through use of color-heavy call out boxes. 
  • The use of color is intentional (not what a graphing application randomly chose); note how red means the same thing in both graphs: total. This consistency is important for easy comprehension.
  • There isn't any extraneous stuff to dilute the audience's attention. Everything that's there is adding informative value.
  • The visual hierarchy of information is clear. The data draws your attention through color; titles are bold so you can't help but know what you're looking at. Axis labels are less emphasized (but still clearly legible) with non-bold font. Sources and gridlines are there to help interpret the data, but are pushed to the background through size and weight so they don't compete visually with the data.
  • Perhaps most importantly, the visual fits well with the article it accompanies. The graphs reinforce the main takeaways from the article and vice versa.
WSJ in general tends to have effective data representations, so if you're looking for more good examples, this is a great source. I still encourage you to maintain a critical eye when looking at any data visualization: observe what works well and what doesn't and try to emulate the effective parts in your own visuals.

Do you have any favorite sources for good data viz?

Saturday, January 28, 2012

tables that make sense

While catching up on some online reading this morning, I came across an article in the Washington Post titled "College Majors and Their Income Potential". I was curious how my brother, a recent grad with degrees in Italian and Political Science, would stack up. Here is the table that was included with the article:


I started to read the table. It looks nice: visually clean, no unnecessary clutter. When I first dove into the detail, I thought it might be ordered by increasing Unemployment Rate: 7.0...7.7...8.2...5.4? Nope. Ok, perhaps it's ordered by increasing Earnings: $32K...$32K...$46K...$33K? Nope. Alphabetical by Major? Nope. Doesn't even seem to be bucketed into degree types, given that Computer and mathematics are up at the top, while Engineering is at the bottom.

The table is not ordered by anything meaningful, as far as I can see. I find this frustrating. Sure, I can use it to find what I'm looking for (I'm guessing the Humanities and liberal arts segment is what my brother's degrees would fall into...highish unemployment rate and lowish salary would help explain why he's happily crafting his art as a bartender at the moment). But if I want to get much more out of it than that, or really understand how a single degree stacks up to the rest, it's difficult. In other words, the table is overly taxing to read: I have to search through to find the lowest and highest values, and it's really hard to figure out the what the ordinal set (which degree ranks 1st, 2nd, 3rd by the column of interest) would look like.

The lesson? When using tables, think about how your audience will use it and order it in a way that will facilitate that. In this case, I would suspect there are two ways the audience would want to use the table: 1) those who will do what I just did: seek out a single degree, then have interest in understanding it in the context of the others and 2) those who are interested in understanding what the top few or bottom few degrees are based on the figures. Some reordering would help both audiences' needs be met. 

In general, think about whether there's an intrinsic order to the segments that you need to leverage. If not, order by the values your audience will be most interested in or based on the point you are trying to make. If high unemployment rate is your message, order your table by decreasing unemployment rates so the ones your audience encounters first - those at the top - help to reinforce the message you are trying to make.

Bottom line: make your table make sense.

Sunday, January 22, 2012

call for examples

I'm starting to work on content for a few upcoming presentations. The topic, of course, is storytelling with data. I find that one of the best ways to teach this is to show real world examples and the typically slight modifications it takes to go from a standard Excel graph to a visual that really conveys what you want it to for your audience.

I'm always looking for new material, so thought I'd put a request out: is there a data visual (graph, table, slide) that you've made or come across that you'd subject to a cole-critique and makeover? If so, please send to me at cole.nussbaumer@gmail.com. Bonus points for including some context on the story you want to convey.


The fine print: Please don't send me any sensitive data that cannot be shared. It's subject to inclusion in a presentation or on my blog. Sending the visual is your tacit approval for this. Submitting a visual does not guarantee that it will be madeover (I'll choose based on the lessons I aim to teach), but if yours is chosen, I'm happy to share the madeover version with you.

I look forward to your submission!

Monday, January 16, 2012

wisdom of crowds: how can we improve this viz?

I am in the exciting process of selling a house. In case you're unsure, that sentence is dripping with sarcasm. While I tend to have fun on the buying side, selling (at least in my recent experience) seems to be all about fixing things and losing money.

Ok, that's enough ranting (almost)...on to the data viz. The broker I'm using apparently has a kiosk in the local mall that is meant to drive traffic to their office. My mild curiosity (and disbelief that this is an effective way to find buyers) prompted me to ask my agent what percentage of sales can be traced back to the kiosk. She sent me the following graphs:



While I can answer my question based on the visuals (4% of transactions appear to be driven by the kiosk - not much, but more than I would have guessed), these graphs seem to be poster children for how not to present data. In my current ranting mood, I could go on and on about what I'd change, but instead I'll try to get over myself and let you join in on the fun: what about the above visuals bothers you the most? Leave a comment. Bonus points for discussing what you'd do differently if you were presenting this information.

Anyone looking to buy a cute rambler in Poulsbo, WA? :-)