Get Your Hands Dirty With Data


The Idea: If the best training is hands-on training … how do we improve our training on data analytics? Teach yourself statistics.

Yikes, you say. Seriously, you say? That’s pretty daunting. Yes, but stay with me:

“… for most people, the gulf between recognizing the importance of data and actually beginning to analyze it is massive. How do those without extensive training in statistics equip themselves with the skills necessary to thrive (or even just survive) in our age of ‘big data’?” Walter Frick writes in a Harvard Business Review blog post.

The 2012 U.S. presidential election elevated the visibility of the statisticians behind the campaign scene, crunching data to predict election outcomes at the most granular level. Nate Silver emerged as the pre-eminent guru in the field. His FiveThirtyEight blog in The New York Times was a must-read for campaign staff and political junkies. The FiveThirtyEight blog recently moved to ESPN and ABC News.

Nate Silver is the man when it comes to data modeling.

HBR’s Frick recently had a conversation with Silver; the main point of their discussion: “Far from counseling that everyone must major in statistics, in the edited conversation he advises students and executives alike to roll up their sleeves — no matter their statistical literacy — and get their hands dirty with data.”

OK, you say: I’m in. But where to begin? How do I increase my statistical literacy?

The Execution: Have you heard of Khan Academy? Its mission: “A free world-class education for anyone anywhere.” If your math and stats brain has gone a bit fuzzy, this is a no-cost, at-your-own-pace way to refocus it. An amazing resource for learners, Khan Academy could be an answer to the age-old internal audit dilemma of how to provide the relevant, just-in-time training our teams — and ourselves — need and demand at a cost that fits into our budgets.

Posted on Oct 2, 2013 by Carolyn Saint

Share This Article:    

  1. Hi Carolyn. It's good to have a blog on auditing at the 'coal face'. While I agree with 'Get your hands dirty with data', I think that there is much that can be done before the need to understand statistics. Although my experience is some years out-of-date, I wonder if the problems have changed? The first problem was always getting your hands on the data to analyze it. What data is available? Where is the data - which computer tables? How can I access/copy the data to analyze it? Whose permission do I need? What analytical program do I use? Ideally a knowledgeable computer auditor is essential to carry out these steps, or if one is not available, a friendly IT department database specialist. I just hope that modern software packages are more helpful answering these questions than they used to be. Once the data is available, there is much that can be done by simple analysis of it: missing data (employees with no social security number, products with no tax code); calculations such as stock to sales (why have we got 200 years' stock of floppy discs at full cost?); expenses (driving around the world three times in a month is impressive). There is also the possibility of using programs which have inbuilt statistics. My department used @RISK to analyze investment decisions. Perhaps I'm out-of-date and everyone uses Computer Assisted Audit Techniques and the next step is to use statistics, but I have a suspicion that auditors may need a course on CAATs and databases before one on statistics.
  1. @David Griffiths:  What's a "coal face?" Agree with your points and don't dispute them however I'm talking about understanding what the computer/excel/ACL software spits out (i.e. what's a median, mean, statistically valid sample, etc.).  The concepts underlying the math.  We also need to be able to ask great questions to even know what data to go after.  And then yes, your experience is still valid--getting the data is difficult.  I just think getting a bettter handle on the theories underlying data analytics that can then be explored via understanding how to use a data modeling tool, learning how to extract and clean a necessary thing. Thanks for your insights!

  1. @David Griffiths:  thank you! (I looked up "at the coal face.")

  1. Sorry Carolyn 'coal face 'is a British expression (somewhat ironic since we have hardly any coal mines left). I can see where you are coming from and agree that understanding what the increasingly sophisticated software spits out is important, particularly the ability to produce a statistically valid sample. My technique of sticking a finger in a pile of invoices is, thankfully, past.
  1. Dear Carolyn,

    Thank you for your article, it is very interesting. I want to share with you that despite more that 150-200 hours of CAAT's training with ACL  as trainer, mock CAATs audit groups,  and with a strong investment in licences, We never could manage real Big Data. Because when Jr. or Sr. want to use/apply  it, Managers and Directors refuse to receive DB in ACL's Format. Another point is that old style managers do not accept this great vulnerability as what it is... our main threat for efficiency and  good audit results. So when Auditors try to aply automated controls on big DB, they did not have good guide from their managers or supervisors in order to get the most of the work. Really it is a big problem for our profession. We could not linked ACL with SAP, despite we bought all ACL programs because Directors never considered this goal as strategic.

    I am not talking about my actual company, I am talking about one of the main carmaker of the industry.

    As Training & Development Manager who had been failing on this goal during 6 long years, it had been ever my professional pain ..



  1. Great stuff, Carolyn.  The Khan Academy knowledge map speaks to me.  -- Jim

  1. Thanks for the comment Jim.  Khan Academy is a world-changer.  It's a great tool for upskilling yourself.  The key remains the individual's personal commitment to spend the time and energy to just be better. 

  1. I believe data analytics is a core competency of every internal auditor. Not just for the IT or specialist but every auditor. The learniing journey will cover these areas. Few will make it but generally it is because they are not properly supervised and/or motivated. 1. Apprentice - you have to set a goal that you need to learn. find a good mentor and maintain a steady pace of using and learning. 2. Generalist - can function and apply data analytics in everyday audit work 3. Wizard - think you know it all as a powerfull position and not really interested in developing staff (fear of competion and losing) 4. Professor - this level is the ideal and it would be great if every audit function had at least one. Interested in collaboration and improvement on a regular basis. May not know all the answers but knows how to get them. If it is not the fool, then try another tool like IDEA!

Leave a Reply