This blog is about test-driven data analysis.

It is written by Nicholas J. Radcliffe, of Stochastic Solutions and the Department of Mathematics at University of Edinburgh.

The ideas have many parents.

First, Sam Rhynas and Simon Brown, who also work at Stochastic Solutions, have been integral to developing the processes that have led to my writing this, and I am discussing and developing these ideas with them constantly.

Secondly, it is work with our clients that has formed the reason for us to think about these things. In particular, we currently work extremely closely with and for Skyscanner, whose use of data and ambitions for this area are practically without limit. Again, without them, this work would not be developing in the same way, and would not have the same internet-scale and internet-speed constraints with which we work daily.

Thirdly, my long-term collaborator, Patrick Surry, has been instrumental in formulating these ideas, and I discuss these topics with him regularly. He is also the co-author of the PMM metadata specification and implementation, which we will soon make available.

More generally, pretty much everyone I talk to about the ideas of test-driven data analysis seems interested by them, and every conversation tends to lead to another idea. I thank everyone listed, and others not listed, for their input.