Main Area and Open Discussion > General Software Discussion

What software will do this? Simple data-mining, finding associations etc.

(1/1)

suleika:
I'm having trouble finding a tool for some basic data analysis.  I want to input discrete lists of items, which will have some overlap between the lists, and then, for any given item, browse its associations (the frequency of list-fellows).  My data won't be very extensive - roughly in the order of 30-50 lists of 20-40 items each. 

An example of the kind of analysis I need (not what I will be using it for) would be lists of cities visited by my friends.  Each friend has a list containing the names of the cities they have visited.  The minimum analysis I would require is that when browsing "Paris" I would be able to see associated cities in order of frequency, and also the names of which friends have visited. 

Further useful information would be about similarity between lists, and an indication of which lists are the more unusual/original and which most "average".  Also, bearing in mind list similarities, which cities tend to be indicators of averageness or originality.

I don't mind if it's not visually represented, though that would be a bonus.

All I've managed to find is expensive software like TouchGraph Navigator and bewildering toolkits like Prefuse.  I keep thinking I must be missing something.  Perhaps language analysis software?  Some clever Excel trick?

steeladept:
If this is in Excel already, it shouldn't be too hard to do this with the analysis pack for Excel.  That said, I am not sure how to do so right now.  I would have to research it just like you are.

tinjaw:
This is just a drive-by posting so not sure if it is exactly what you are looking for, but you may want to view http://www.slideshare.net/noahflower/overview-of-network-analysis-platforms. I have it bookmarked for later reading.

suleika:
Thanks, both of you.  I realise, trying to use some of the software referenced on the slide show, that I need to learn a little about graph theory etc and find out what corresponds to the type of data I'm trying to analyse.  It's all very interesting.   

PPLandry:
Excel's pivot table will do that for you (Office 2000 and up)

The Excel data would look like this:
Person    |  City
John Doe |  Paris
John Doe |  London
Mary       |  London

etc.

InfoQube would do it too:

* John Doe

* Paris
* London
* Mary

* LondonAnd then use the build-in pivot table

Navigation

[0] Message Index