Month: September 2011

What took so long for viz?

Visualized data seems as natural as eating and sleeping, doesn’t it? Yet the first economic time-series wasn’t plotted until 1786, according to our patriarch of viz Edward Tufte in his 1983 book Visual Display of Quantitative Information.

What took so long? I suppose humanity really did suffer from lack of an Excel chart wizard. People had been making maps for centuries, but apparently no one had made the leap from maps to abstract quantities like time and money. That wasn’t so easy to do, after all. You can’t just scratch numbers in your clay, you have to think about it first.

It all comes back to what’s probably already a cliche: simple is hard.

Still, what took so long?

Survey’s closed, results coming

This afternoon, I finally closed the long-running survey of “those who analyze data.” The results are seeping in to Datadoodle headquarters. I’ll release them in stages over the next two months: first, highlights, then more highlights, and finally a preview report and a final report.

I opened it in mid-February this year. It has 221 responses. I’d say that’s a good number from a small platform like this weblog in this survey-saturated industry.

In case you didn’t realize it, I’m no data analyst myself. My stint in market research that ended more than 10 years ago entailed little data grooming, reshaping, or cleansing. Someone did it for me, and only then would I paw through it. But today, it’s all up to me — me and my handful of data-savvy volunteers.

I’ll identify and thank each one publicly as things progress, along with a few people who helped promote the survey.

The very first look will probably come in my columns at BI This Week (TDWI) and Information Management. After that, I’ll issue a report, first in a preview edition for respondents who asked for it and a few others.

Just watch this space for links.

Everybody talks about collaboration

Everyone’s talking about collaboration — but what culture and tools does it take to succeed? Six principles emerged from a discussion among experts at the Pacific Northwest BI Summit in July.

Even among this moderate crowd, an old tension showed itself. BI has always had what I’ve thought of as “data police,” those who stress data quality, security, and process at the expense of natural workflows. On the other end, “data libertarians” have used self-service tools to wrest control of their data analysis and data itself — an approach that has made more sense to me. If BI is about business, why leave it up to a cabal of technicians?

Read all about it in my BI This Week column. It appears tonight. Here.