Comments on: Exploratory Data Analysis http://blogs.cetis.org.uk/adam/2012/05/18/exploratory-data-analysis/ Cetis Blogs Wed, 07 Jan 2015 09:19:39 +0000 hourly 1 http://wordpress.org/?v=4.1.22 By: More Tukey Gems | OUseful.Info, the blog... http://blogs.cetis.org.uk/adam/2012/05/18/exploratory-data-analysis/#comment-27702 Fri, 24 Oct 2014 15:19:15 +0000 http://blogs.cetis.org.uk/adam/?p=542#comment-27702 […] a half quote by Adam Cooper in his SoLAR flare talk today, elucidated in his blog post Exploratory Data Analysis, I am lead to a talk by John Tukey – The Technical Tools of Statistics – read at the […]

]]> By: David Kay http://blogs.cetis.org.uk/adam/2012/05/18/exploratory-data-analysis/#comment-99 Sun, 03 Jun 2012 12:47:41 +0000 http://blogs.cetis.org.uk/adam/?p=542#comment-99 Bravo – I really agree with the spirit of this post!

I like the way Tony Hirst talks about allowing the data to tell its story. Such data narratives may contain ‘secrets’ that can be more readily oped up through a variety of visualisation tools.

The key is not just being open to such approaches but also preserving enough data to support them, as opposed to aggregating or otherwise limiting the data preserved to that which aligns with preconceived lines of enquiry.

]]> By: Adam Cooper http://blogs.cetis.org.uk/adam/2012/05/18/exploratory-data-analysis/#comment-98 Mon, 21 May 2012 16:33:43 +0000 http://blogs.cetis.org.uk/adam/?p=542#comment-98 Tavis-
I think it is a big issue. And I think “quality” can cover a multitude of sins so in an ideal world you’d want to know and carry forward the information about the context of collection, through the storage, …. Its not just typos and missing data (or what was done to accommodate missing data a.k.a “imputation”); its easy to lose track of what data means. Ask me what I earned last year. Ask my employer. Check my bank account… Ask HMRC. How many different answers do you want?

As for some of the linked data out there: far too much scraped HTML and guesswork concealed inside very-authoritative-looking RDF!

“Not all data is equal…” Definitely.

Cheers, Adam

]]> By: Tavis Reddick http://blogs.cetis.org.uk/adam/2012/05/18/exploratory-data-analysis/#comment-97 Mon, 21 May 2012 13:12:13 +0000 http://blogs.cetis.org.uk/adam/?p=542#comment-97 Is provenance also an issue? If not all data is equal, and some sources are more authoritative than others… Although maybe that belongs to a larger issue of data quality and assigning probabilities/weights/degrees of confidence?

I was once on a work placement as a database administrator where data from event attendee registration cards was mixed with more reliable data sources, and provenance mattered. Possibly even more so in the case of linked data.

]]>