• Subscribe in Bloglines
  • Add to netvibes
  • Subscribe in NewsGator Online
  • Add to Google
  • Instant Cognition Feed for Yahoo!
  • Add to Microsoft Live
  • Get updates in your Inbox:



Instant Cognition Feed

www.flickr.com

Archive for July, 2006

Sharing the Wealth: Random Excel Tip #2

Monday, July 31st, 2006

The guys over at Juice Analytics are at it again.

It’s another one of those ‘doh!’ moments where I say: “Why didn’t I think of that?” Basically, using a simple function in Excel (Rept, aka repeat) which allows you to repeat any piece of text N number of times, you can create simple, in-cell horizontal bar charts. I’ll let you go read their post as they explain it quite well and thoroughly.

All I have to say is, no need to wait for Excel 2007 and their less than accurate in-cell gradients - you can be doing this right now and it’s not only a better solution but also is extremely reminiscent of a sparkline.

From the How Cool Is That? Files

Wednesday, July 26th, 2006

Via Live Science, Technovelgy and some Japanese blog that I can’t read comes this very cool gewgaw.

AMOEBA Tank Forms Letters Today, Aliens From ‘Abyss’ Tomorrow - Create letters with standing waves; will it be possible to create effects like the water aliens in Abyss sometime soon?

Talk about an amazing and inventive visualization…

Photo Source: FujiSankei Business i on the web

Creating relief images in water - that’s SO COOL!

Stephen Few is Looking for a Few Bad Charts

Tuesday, July 25th, 2006

It’s Oh So Punny Post Title Day. :)
None the less, Stephen has started a series of posts under the title of ‘The Graph Hall of Shame’

You can e-mail your personal gems for nomination to sfew-at-perceptualedge-dot-com

If Stephen picks one of your submissions, we get the added bonus of he and other visualization aficianados commenting on all the issues and maybe even suggesting a tip or three to resolve those problems.

Tables Are Data Visualization Tools Too

Tuesday, July 25th, 2006

Data tables often get a bad rap in the data visualization world. I think this is because your default table whether in Excel or some other tool is amazingly boring and can do a great job of turning off the brain of you report users. However, it’s important to note that even the gurus like Edward Tufte and Stephen Few find visualization solutions with tables. Here’s a perfect example, illustrated in Perceptual Edge’s (Stephen Few) Gallery.

In it, Mr. Few takes a poorly designed polar plot and organizes the data succinctly into a table that would allow the end user to quickly and easily make the appropriate comparison.

Here are a few of the reasons I think people think tables don’t work or tables fail:

  1. Reliance on the default appearance from the application
  2. They’re not ‘cool’, ’sexy’, ’slick’, or like charts, graphs, and maps are
  3. Table junk - just like chart junk, but in a table and in either venue the junk can ruin it
  4. Overuse of either horizontal gridlines, vertical gridlines or both
  5. Tables have two masters: visualization (design) rules and text (reading) rules because they are text-heavy users will, in part, have an expectation of reading like they would read a book, article, paper, etc

I happen to use tables all the time in my work so here are some of my tips for designing tables in Excel

  • Don’t rely on the default gridlines. Once you have your data organized in the table, go to Tools > Options > View and take the check mark of ‘gridlines’
  • Always have column headers and row headers (i.e. describe what the metric is labelled correctly, typically in the column header although your mileage may very, and what is being measured is in the row header)
  • DO NOT justify the text. If the text is all justified, this diffuses the focus and makes it difficult for users to pay attention - align left, center or right as appropriate. In other words, ragged-edge text is easier to read
  • Clearly seperate row and column headers from the data - this can be done using borders, cell color or both - but be careful about it, less is more
  • Use borders sparingly, especially horizontal borders. Anything that over-enforces the regularness of the table should be avoided - most users only need a hint of structure in order for them to internalize it, this is the same rule we apply in chart design when removing or de-emphazing axis gridlines
  • Use an appropriate sized font - probably 10pt or higher, but it depends
  • Whatever you do, make sure that the focus is on the data, if users are paying more attention to the row and column headers, to the font choices, to the gridlines they aren’t getting what’s important. Give your draft table to a buddy at work and ask them if they can read it easily, it’s a great test to see if you are on the right track.

Welcome to the new Instant Cognition blog

Sunday, July 23rd, 2006

The past couple of weeks, I’ve been working hard to get a new version of Instant Cognition up and running.

Finally, I think it’s ready. As I am also sure that it has some undocumented ‘features’, please let me know when you encounter one.

If you had been linking to instantcognition.blogspot.com, please update your links to point here as I will be decomissioning the blogspot site sometime in the near future.

Enjoy the new site, and of course, please leave a comment if you have a thought you’d like to share.

-Clint