Analysis Tools for Everyone

by Sandy on October 29, 2008

in General / Tips

I’ve introduced you to Stephen Few here previously. He specializes in the visual side of business intelligence, and always makes for an interesting read with both his writing style and his content.

Yesterday Stephen wrote about how he and the people at Tableau Software really feel that the need for visual analysis tools is becoming universal; no longer the exclusive domain of analysts, but of anyone who needs to understand and examine information.

There’s that common message once again.

Of course I’m a bit biased, but you know, if you took out the words “visual” and “visualization” and the references to Tableau, you’d swear that Mr. Few was discussing how anyone that uses data should excel with Monarch. ;-)

{ 2 comments… read them below or add one }

Grant October 31, 2008 at 6:01 pm

Sandy,

You know I have some real reservations about the way mainly visually delivered information can be detrimental to understanding, interpreting and decision making. There is nothing wrong with a visual representation per se, indeed it can be very powerful. And that is the problem.

Introduce some chartmanship and you can readily mislead people with pretty graphic. Selective scales for example. Or colours chosen for their ability to influence graphic impressions as interpreted in the brain.

People may make their decisions based on their expectation of the graphic fitting ‘the usual conventions’. No data checking required.

So lets say, for example, that you are using the ‘standard’ traffic light colours that we all know and love and that your charting application can be set to create a line graph where the width of the line varies (and with it the impression that is leaves) according to the values being reported. We might expect the convention to be that Green is good, Red is bad and a wider line means GOOD of its a green line but BAD if its a red line.

Now we set the parameters for the graph to be GOOD = RED, BAD = GREEN and smaller number = WIDER line (for example) or vice versa depending on what we are trying to present. The ‘at a glance’ graphic could insinuate a completely different meaning than the raw data indicates. Few would know.

There are great benefits but also huge challenges here unless people can be persuaded to follow ethical practices. In our modern world that seems to be a big ask.

Grant

Sandy October 31, 2008 at 9:52 pm

You’re correct. Any powerful tool or even technique can be misleading.

Even Excel can be misused.

If a Monarch model is prepared even slightly incorrectly, for instance by not setting fields wide enough to capture widely spaced negative signs, it could provide results that could be construed as misleading, even if it was simply an oversight.

Even just distorting the aspect ratio of a chart can mislead the reader.

Being an informed reader is perhaps just as important as being skilled in the preparation of visual analysis, as such a reader would (hopefully) be aware of the tell-tale signs that someone is trying to pull the wool over their eyes.

Of course, having a solid understanding of the topic being analyzed is invaluable.

Leave a Comment

*

Previous post:

Next post:

Copyright © 2007 - 2014 Excel with Monarch Training and Services. All rights reserved. Privacy Policy
Microsoft Excel™ Microsoft Corporation. Monarch and Monarch Pro™ Datawatch Corporation.