I found something at the weekend that just blew me away in terms of impact. Something that communicates both geographic location and volume in a moving time-series.
This is the growth of Target stores in the US since 1962. It is a visual feast, watch it here.
It follows an equally impressive visualisation of the growth of Walmart in the same period.
The Walmart visualisation can be found here.
What I would like to do is plot student activity across Australia and indeed around the globe for an Academic Year, a Teaching Period or maybe even a day.
The activity could be an enquiry, an admission application, an enrolment or maybe smaller-grained interactions like an LMS login.
Does anyone profess to know how to do this? Nathan generously lets you download his code here but the technology is not something I’m familiar with. He’s posted a comment saying
Actually, I use Flex builder 3 for all of this since it’s all actionscript. It’s available for free 30-day trial on Adobe, but even better, it’s free for students: http://flexregistration.com
I haven’t used Flash, but I’m pretty sure there’s an equivalent that you can use - like import flash.effects.easing… um, try googling flash easing maybe?
So come on industrious students, I’m sure we can work out some kind of casual employment deal if you think you can help out. We’ve got plenty of data…
This is very telling communication!
Note how the Target example burgeons through the south, then the west coast with quite wide leaps between areas of activity. It shows quite a lag (more than 30 years!) before suddenly spreading into the northeast.
And the Walmart visualisation shows a very connected pattern of growth concentrated for a long time in an ever-widening pool around the origin before leaping further afield.
The end states of each visualisation seem to map population distribution (see http://earthobservatory.nasa.gov/IOTD/view.php?id=7052).
It would be truly interesting to see at least one other dimension in each of these visualisations. Population growth is an obvious choice but what else would be informative from a BI point of view? For example, Target clearly didn’t map population distribution in its opening of stores: they didn’t get into the densely populated northeastern states until the mid-90s so purely demographic factors are not the whole explanation.
Switching to the Australian higher education context, it would be great to visualise population growth, the establishment of universities and CAEs and then the proliferation of private providers. But maybe it also would be informative to look at the distribution over time of students against the distribution of institutions. For universities we would only need 38 colours on the map!
Seriously, there is a challenge to represent comparative data in ways that aren’t confusing but lets start simple and build to the more complex. I hope creative people will take up Rob’s challenge.
And for those who want to get started, here’s O’Reilly’s book: http://www.adobe.com/devnet/flex/pdfs/getting_started_with_Flex3.pdf
Nice post thanks for sharing
thanks rick nelson