Do You Suffer From Analytic Personality Disorder (APD)?

Dominic Vincent Ligot, Teradata.

Are your company executives crazy about analytics? As in LITERALLY CRAZY?

Since we're entering Halloween season again, I thought it apt to talk about some of the most bizarre personality glitches I have noticed in the recent year – it appears the big data era has brought some possibly really scary disorders upon us!

Disclaimer: I have NO background in psychology. Nonetheless, inspired by the APA’s DSM (Diagnostic and Statistical Manual of Mental Disorders, now on its fifth revision), to (sort-of) qualify as a “personality disorder,” the cited pattern should not be due to direct physiological effects of a medical or chemical substance (so far as I can tell) or a medical condition (like head trauma – much as I would have liked to “bonk” these people hard on the head sometimes). Then again, the cases below are anecdotal ones I picked out simply because the people who exhibit these behaviors are just asking for it.

Auto-esotericism/Rube-Goldberg’s Disease

A pervasive tendency to over-complicate a solution, sometimes indicated by:

• To the tune of: an insatiable itch to deploy elaborate multi-class polynomial regression models, random forests, or neural networks, to tackle simple problems that a moving average could have solved (even if realizing this after the fact).
• Pervasive self-justification: Always taking the voice of popular author or influencer, common candidates are Tom Davenport, Clayton Christensen, Guy Kawasaki, or the classic Steve Jobs (at Stanford speech) as justification for trying an overly complex and untested strategy.

Codophilia/Codophobia

• Codophilia: a condition indicating a pervasive need to render anything in complex code like SAS, SQL, Java, Python, R, or even C… sharp – even if simpler and less manual methods are available. Tendency to over indent, punctuate sentences, and the use of [brackets], {braces}, and (parenthesis) in a paragraph to over-compartamentalize one’s thoughts.
• Codophobia: The opposing behavior indicates an reflexive repulsion to anything code-related (e.g unconsciously vomits at the mere sight of a semi-colon, hash, backslash, and especially the word SELECT). Will prefer an overly elaborate GUI with twenty-five widgets and thirty sub-menus to a simple one-statement macro.

Credentialism/Title-fetish

• A pervasive, possibly fatal obsession with degrees and job titles. Some Data Scientists have been quoted as likely to murder anyone who accidentally calls them Data Analysts, while BI Analysts usually don’t seem to mind being mistaken for Data Engineers, but never as ETL Developers who in turn are almost always confused with DBAs. Most have difficulty filling up online forms due to obsession with alphabet soups: MsC, MBA, PhD, ABCD, MD, JD, PMP, EFGH, ITIL, MSCE, IJKLM, d - All of the above… drat.
• A mirror behavior amongst the C-Suite: CTOs hate to be called CIOs, the CMOs will rebrand themselves CDOs – but wondering which “D” is more prestigious: Digital or Data, but appear to successfully dodge some obviously risque job titles like Senior Transformation Director (STD), or Chief Data Scientists (CDS) who don't exist in the investment banking industry for some reason and Principal Managing Scientists (PMS) who might otherwise have a cheery disposition.

Data-mania / Data-mentia

• A pervasive pattern of grandiose thinking and tendency to proclaim data as the panacea for all the company’s ills. A common condition observed amongst mid to senior levels, also associated with recent attendance to a big data or vendor conference. In the case of serial conference goers, they predictably go through the 5 recurring stages of big data mania: Denial (of course we are digital), Anger (what do you mean we’ve blown budget), Bargaining (can’t we get even a single report to work), Depression (they said Hadoop can do everything) and Acceptance (there’s always next year’s conference).
• Other indications include a long list of failed and half-finished data initiatives. The road to hell so to speak.
• A variant of data-mentia is a pattern of obsessive, almost cultish, veneration of data. Oft saying: data knows all - only data can lead you to truth. Afflicted analysts... or scientists rather - drunk on confirmation bias will draw out even the most anecdotal "insight" from random data points and spurious correlations to defy common intuition - (since my birth I have not experienced any form of death, therefore evidence supports that I am immortal).

Humanisnomerism

The excessive push-back to Data-mania: A pervasive skepticism and denial of anything data-driven – and obsessive prioritization of human intuition and reasoning, ESPECIALLY WHEN evidence seems to directly fly against it. Other common indications:

• Secretly sabotaging (or wishing / wanting to sabotage) computers, algorithms, even data feeds – if only to prove the risks of relying on automated decisions. Loves pointing out the security risks of cloud computing – even when most (or at least a higher percentage of) examples of significant data breaches are on non-cloud on-premises systems, or really a breakdown in ethical standards of employees rather than failure of a hosted data center or the like.
• Celebrates errors committed by a classification or prediction engine even if this accounts for just a 4-sigma chance likelihood, ignoring 99.99% of confirmatory evidence. (Hey black swans right?). Also an obscene fascination with Facebook’s facial recognition failures (mistakes an elbow for your friend’s face), and despising their smartphone’s autocorrect or when Waze recommends a U-Turn twice in a row.

Open-source-miscuity/Toolchosis/Vendor-obsession

The pervasive tendency to prioritize a particular tool, vendor, or open-source solution, over and above its logical and technological merit.

• Open-source-miscuity is the pervasive preference for any open-source toolset, Apache project, even abandonware over commercial tools - regardless of effectiveness or efficiency (and also indicated by an inordinate amount of time spent on google, git-hub, stackexchange, stackoverflow, or similar forums trying to find an obscure answer to an analytic problem).
• Toolchosis is a specific preference for a specific tool, program, or application usually from tradition or personal experience. Vendor-obsession is a wider preference for commercially-branded products from a particular publisher or vendor regardless of usefulness (and usually encouraged by vendor freebies like coffee cups, pens, notebooks, and black jackets).
• Cases of vendor-obsession markedly skewed towards those with prominently-blue 3-letter logos. Toolchosis often observed among those who have been using MS Excel for years then suddenly exposed to fancy bubble charts. Open-source-miscuity is also strongly correlated to codophilia, credentialism, and auto-esotericism – sometimes it’s tempting to think they are the same disorder.
• Frequent faux-commercialism associated with these behaviors (i.e. open-source is free, so we are saving costs, vendor support is cheaper if we buy full stack, cloud requires no hardware so we are saving maintenance, self-service BI will increase our overall productivity, blah blah blah), ironically without any business case or financial evidence - even in a few cases actually working to suppress any cost-of-ownership audits.

Can you cure APDs? Unfortunately, like snoring – APDs are seriously difficult to self-diagnose – you need someone else to tell you that you are suffering from it. And once identified, sometimes you wished you could hand out bitter pills like these:

• Auto-esotericism – Take them to task against simpler solutions and collect a fine whenever they unconsciously quote a dead mathematician or big data influencer.
• Codophilia/Codophobia – deliberately (even if painfully) have a philiac and phobiac switch roles for a week or two then have them observe each other's results. Often the fixes and solutions pursued by one will be an eye-openers to the other.
• Credentialism – call everyone data janitors for a week (because that’s what they probably do). See if the reality sinks in.
• Datamania – Quite hard to intervene – since these cases often involve very senior people. Board involvement can be sought to impose an embargo on all big data conferences for at least a year. Also have these senior executives immerse with the staff (read: data janitors) for a few days, for them to see first-hand the massive implications of each of their random statements on the workload and morale of the people.
• Humanisnomerism – deprive them of phones and the internet for a few days. See where they end up.
• Open-source-miscuity – similar to the humanisnomerists, deprive them of google for a few days. See how much they accomplish.
• Vendor-obsession/Toolchosis – Hard to extricate people from this. Maybe deduct the license costs of their favorite vendor or tools from their salaries and see if their behavior changes.

A final warning: attempts to explicitly treat APDs – could result in much grief, not to mention reduced finances for individuals and companies concerned. At least, I can imagine it could.

Meantime, tread with care.

Big data has made some really crazy people out there.

Happy Halloween.

Bio: Dominic Vincent Ligot helps organizations gain competitive advantage through data and analytics. Great analytics requires good data and great people - I’m passionate about both. I am also usually PC, but sometimes it pays to be a little irreverent to “big data wackos” these days (in my desperate attempt not to become one myself - as Nietzche supposedly said). The DSM is also one of the most fascinating textbooks I have ever tried to read.

Also, don't take any of the above too seriously.

The Hannibal Lecter image is copyright owned by Orion Pictures. Unattributed asylum image found on Pinterest.

Original. Reposted with permission.

Related:

• Still Searching for ROI in Big Data Analytics? You’re Not Alone!
• Do Multipliers Trump Big Data Analytics?
• How top companies use data to make wise decisions