All types of web objects can trigger conversations:
- books at Amazon or LibraryThing
- pictures at Flickr or Picasa
- links at del.icio.us
- questions at Yahoo answers
- blogs – at Technorati or Twingly
- news articles and blog posts – all over the place
So why not statistical data sets?
Shared data sets
Being a statistician, I often try to understand what is happening in libraries by looking at the statistics. This usually means that I select and rework official statistics.
Data as such are dumb. To get knowledge from statistics, we must first formulate specific questions – and then organize the data in such a way that they can answer those questions.
In my small collection of Norwegian data sets, there is one spreadsheet showing Book loans per capita in all Norwegian municipalities for the years 2001-2007. The municipalities are diveded by size intoo five groups. The data come from official statistics. But the particular organization into years and size groups is my own.
I set up the spreadsheet to understand what is happening to book loans over time. I used the size groups because big urban libraries tend to change more rapidly than small rural ones.
These data can be used in many other ways. We can look at single libraries or compare libraries within the same size group. Iw we add non-library data, we can try to find relationships between book lending, on the one hand, and – say – the presence or absence of book stores, on the other.
Once the data set has been published as a web object, I could add a comment button – and we could start a discussion linked to this particular spreadsheet.
In fact, I’ll do that. I do not expext a rush of statistical comments. But I want the possibility to be available.