By all means experiment, but please do so intelligently

Experimentation and Design of Experiments

  

Experimentation is a ‘new’ buzzword, some even claiming “Big Data” demands experimentation.

I agree, and disagree. 

The need for experimentation has always been here – why would you launch a full scale campaign costing hundreds of thousands when you can spend some time refining and enhancing your offer and targeting?

It’s just that, with “Big Data”, data is available at a more granular level and hence we are able to detect more subtle events and sequences of events that can be seen as precursors to various courses of action. Today, organisations can quite easily know where their customers are, and allied with their recent behaviour, can create tailor-made offers. If I just had lunch at a fried chicken restaurant, and my bank has an offer at a yogurt bar nearby, it might be the exact thing to balance my tummy. Hence the bank could send me an offer for the yogurt bar.

Technology has helped in 2 further ways. Firstly, not only is “Big Data” available for analysis at a very granular level, but we are able to do so without having a Computer Science Degree. Hence the number of potential experiments that can be attempted has increased. Secondly, technology has also increased the ability to execute these experiments; apps, text messages are all avenues where we can be reached.

But what, to me, is critical is the design of the experiments. Just like with “Big Data” analytics, while technology has put the ability to analyse data in the hands of many, there are fundamentals that put probability on your side. Proper experimental design is critical and the costs of not doing so can be catastrophic.

The articles linked below highlight some failures in the medical field that have huge implications and that could be avoided simply with proper experimental design. The sad part is that the failures are so basic that a little thought would have avoided them:

                Ensure large enough a sample to conduct experiments (power tests)

                Do not bias the experiment by selecting groups to affect success probabilities 

(random selection)

                Do not bias the experiment by your own behaviour (double blind tests)

http://www.theguardian.com/science/2015/apr/18/animal-lives-wasted-in-drugs-safety-tests

http://www.cbc.ca/news/health/flaws-in-medical-experiments-on-animals-a-waste-1.1407577

By all means experiment, but please do so intelligently.

How governance should take advantage of technology/"Big Data" and stop making excuses.

How governance should take advantage of technology/"Big Data" and stop making excuses.
( Why the 30000foot defense shouldn't be allowed in this world of "Big Data")

“Big Data” has been touted as the panacea, personally I think it is important to understand what is worth doing and what is not, to discuss what we should allow to be done, and what not, and be allowed to change our minds as time goes by as we see the consequences of our decisions.

One topic where, I believe, “Big Data” is very useful is in that of governance. 

The amount of data that is captured within an organization, together with the technologies that democratise the ability to extract or pursue meaning in a mass of data should make it impossible for a governance body to claim the 30000 foot defense: “we are too high up to have known”. No, sorry; you should have known (especially when you are paid very handsomely to simply ensure governance).

What can big data and appropriate visualisations allow you to see? Everything down to the minute details should you choose to:

With text mining you can scan through megabytes of documents and instantly get the gist of it all, allowing you to query further (yes, you can have ‘experts’ to do the querying if you want, but you cannot say you had no idea). For example the summarization of the Senate report on torture:

All you have to do when you see something you don’t find kosher is to ask: “tell me more about this”, and you drill down further.

Or if you want to understand how people’s or organisations’ behaviour varied, you could visually see how they differ (the example refers to an earlier piece on whiskies rather than people or organisations, but the idea is the same, comparing attributes across individuals):

And with some mouse work:

Ask “how does Mr. GlenFiddich differ from the rest?”.

Or what do the relationships between my clients/relationship managers look like?

And what is this interesting relationship in the centre?

In sum, with “Big Data”, there’s very little going on, especially within an organization, that can be hidden from people whose mandate is to ensure everything goes according to expected standards and values; unusual behaviour leaves traces that are so visible; you don’t need to be a rocket scientist to uncover them.

This blog post was inspired by: “http://www.theguardian.com/commentisfree/2015/mar/13/bankers-hsbc-rona-fairhead-firewall”

Visual Text Mining: the starfish of torture

There is so much information in text that it is sometimes difficult to choose what to read.

Text mining is a great way of extracting the main ideas from a body of text, but I think that it has to be allied with good visualisation, to capture and make it easy for anyone to understand the main ideas of a body of text.

The US Senate released a report on torture a few months ago, and I tried to gather the main conclusions and express them visually.

The summary is as follows, the starfish of torture:

I have created a deck of slides that illustrates the power of visualisation in bringing text mining to life and capturing ideas.