it would be intersting to get the data in cross tabulated (array) format.
for exemple : class vs race
this format allows a mapping trough factorial correspondence analysis and gives a fast graphic idea of pairing.
it can either be displayed in the quickstats or maybe in the query tool.
if you want to have a go, just send me data in an Array (like global class vs race @ 50-60) and I'll process it for you and return a graph that will certainly show very interesting things.
// edit I did a first mapping using your quickstats on the serveur type vs. class. I have no means to upload it on the web, but I can mail it to you if you will (right now it's 20kO in excel format, but I can put it on a JPEG)
cross tabs => class/race mapping
this is fine in the principle, though I have a problem processing the Race Class Stats :Rollie wrote:Check out the Race Class stats
the column for druid is missing data and there is a strange (german probably) "Untoter" line.
yep you need to be careful with the class names in the Euro version ! I'd say the simplest thing is to put everything in english before processing and getting it back after if needed.
yep a graphic mapping can be fine making the data more readable, the drawback is you need to give some explanation on "how to read", but it's clearer than plenty of percentages.I need to do something to clean up the quickstat page as it is quickly becoming unwieldly!
now maybe you should publish all the data all the time, because it's not quickstats anymore
Because once you get on that track, you'll always want to make more and more. For me you can enrich the quisckstats a bit, but them you need to make a dedicated page at a given date for a dedicated study of this or that .... e.g. a mapping can allow for the study of changes too, like "what are the rising classes" (since a certain date).... these little studies, released from time to time, can help keep the census on top of debates.
as soon as I get the data I'll produce the mappings and send them to you for review.