Skepticizzle & Wisdom!

A Great Analyst’s Best Friends: Skepticizzle & Wisdom!

Original article:

Here’s suttin’ blingin I’ve observed up in mah experience up in hustlin wit data, n’ changin crews wit ideas: Great Analysts is always skeptical. It aint nuthin but tha nick nack patty wack, I still gots tha bigger sack. Deeply so.

This was always true, of course. But, it has become mission critical over tha last few muthafuckin years as tha depth, breadth, quantitizzle n’ every last muthafuckin other dimension you could apply ta data has simply blew up like a muthafucka. There is too much data. There is a fuckin shitload of tables/charts/”insights” bein rammed down yo’ throat. There has been a explosion of “smart-ass muthafuckas.”

If yo ass aint skeptical, yo ass is goin ta take a thugged-out dirtnap (from a professionizzle perspective).

And, yet… Yo ass can’t be paralyzed by skepticizzle fo’ realz. At some point, you gotta jump. Or, yo ass is dead (again, professionally).

Let’s do dis post up in two pieces.

First, a plea ta be skeptical, of every last muthafuckin thang n’ everybody, illustrated rockin a example from one of da most thugged-out bigged up sourcez of data up there, so peek-a-boo, clear tha way, I be comin’ thru fo’sho. Followed, by lyrics on gettin ta a thugged-out decision rather than what tha fuck happens ta skanky analysts: paralysis.

Yo, second, as we is on tha topic of pimped out analysts, I wanna share how tha fuck ta recognize dat you might be one, from a macro perspective, and, if yo ass is, or is not, what’s yo’ value ta yo’ company.

Yo, surely, yo ass is intrigued dawwwg!

#1A: Skepticizzle is yo’ BFF.

I saw these two numbers presented tha other day: 42% of online shoppers use vizzle fo’ pre-purchase research. 64% use YallTube ta find shizzle.

As soon as I heard them, I knew they was horse-manure.

Da source of skepticizzle was simple, neither number is legit fo’ me – n’ I’m up in a place, wit people, whoz ass is da most thugged-out connected playas on tha hood wit mo’ devices ta do dis type of research if dat shiznit was true. I stood up. Did two thangs. I axed tha 100 or so playas up in tha room if either of these two numbers was even close ta reflectin they reality, one thug raised they hand. Y’all KNOW dat shit, muthafucka! Then, I axed fo’ tha source of data fo’ realz. A 2014 AOL report n’ a online survey wit n=600.

Dat shiznit was horse-manure.

Yet, they was bein presented as facts on a tablet handed ta Moses.

Yo ass might not yet have tha experience ta know if a number is legit or not, like yo ass is evolving. But, if you actively invest up in yo’ ejaculation, awareness, bein horny ta always wanna dig just a lil deeper, you’ll git there up in no time.

For example, you might read NADA say this: “85% of hustlas made up they mind ta purchase a hoopty before they left they house.” Yo crazy-ass skepticizzle radar should go beep, Beep, BEep, BEEp, BEEP, n’ you should stop n’ dig dat shit. Well shiiiit, it do not matta how tha fuck big-ass NADA be n’ how tha fuck nuff analysts they have – cuz acceptin imprecise shiznit will cause you ta make game-limitin recommendations.

Here’s a pimped out example of hoppin on tha skepticizzle train right away.

[Update: As Mauro Avello points up in his comment, dis could be a April Fool’s joke by tha crew at Da Economist. I did notice tha date when I freestyled dis post, n’ read tha comments n’ follow ups, n’ did not peep anythang dat indicated ta me dat dat shiznit was a joke yo. Hence dat shiznit was used below. Da lessons you’ll learn still stand yo, but please do be open ta tha possibilitizzle dat dis was just a joke.]

Da eva straight-up dope data viz crew at Da Economist had a irresistible link: Ice Cream n’ IQ.

Hard not ta click on that, right?

It be a gangbangin’ finger-lickin’ dirty-ass short article containin a line chart plottin ice cream consumption on tha x-axis n’ tha mean score on PISA readin scale…

economist ice cream-PISA scores

THE DATA TEAM (that’s whoz ass tha article is credited to) go on straight-up seriously ta share dat mo’ ice cream smokin might be tha solution ta skanky hustla score. They dutifully compare tha Aussies n’ tha Finns, commend tha Canucks n’ crap on tha Peruvians.

Yo, so. Yo ass is tha smart-ass Analyst.

Yo crazy-ass first skepticizzle flag should be: Da title of tha article say IQ, do PISA scores measure IQ, biatch? Quick Gizoogle search. They do not.

Red flag.

Yo crazy-ass second skepticizzle flag should be: Look fo’ thangs up in tha data-set dat disprove tha summary statement. Notice Hong Kong, Singapore n’ it’s neighbors have straight-up high PISA scores, yet straight-up low ice cream consumption.

Red flag.

Yo crazy-ass third skepticizzle flag (for a smart-ass Analyst, probably dis is tha straight-up original gangsta one) should be tha perennial favorite: Correlation do not imply causation!

Yo ass pour over tha article fo’ signs dat dis simple rule aint broken. I aint talkin’ bout chicken n’ gravy biatch. Is there anythang dat shows they looked tha fuck into causation, biatch? No.

Giant red flag.

And at dis point, a tiny part of y’all also took a dirt nap cuz you do so ludd THE DATA TEAM all up in tha Economist.

To aiiight playas (non-Analysts), dis graph n’ article looks legit fo’ realz. Afta all dis be a reputable joint n’ it aint nuthin but a reputable crew. Oh, n’ look there be a red line, what tha fuck be lookin like a funky-ass believable distribution, n’ a R-squared hommie! Most aiiight playas will take dis as truth (and at least 67 of dem will proceed ta comment on tha article n’ have fun).

Yo ass should not.

Da thang dat should go all up in yo’ head is… Causation. I aint talkin’ bout chicken n’ gravy biatch. What could cause this?

Here’s one hypothesis: Muthafuckas whoz ass straight-up care bout they lil playas ejaculationizzle accomplishments come from crews dat tend ta have muthafathas whoz ass is a lil bit well off – middle class -, they can focus on tha kids. These crews probably reward ejaculationizzle accomplishments, n’ you can put dat on yo’ toast. Da reward of chizzle tendz ta be ice cream!

Remember, it’s a hypothesis. We can go look fo’ data. If it turns up ta be true… It aint ice cream consumption dat is tha reason fo’ tha performizzle scores, it is tha fact dat crews tend ta git a cold-ass lil certain income. Or, dat they tend ta have structured work time, which gives muthafathas free time ta focus on how tha fuck they lil playas is bustin vis-à-vis ejaculation.

There could be a fuckin shitload of other thangs. Weather n’ shit. Number of dem hoes up in tha ghetto. Longitude. Number of lil pimp workers. Crime fo’ realz. Anythang straight-up.

Look fo’ causation. I aint talkin’ bout chicken n’ gravy biatch. No causation means… data crime against humanity.

Let’s brang dis baby home, one mo’ example, dis one a lil’ bit mo’ fun.

There be a mad tight correlation between tha amount of US bustin on science n’ suicides by hanging… r-squared of 0.997…

hangin suicides-US science spending

If yo ass is wit me thus far, yo ass is beatboxin dat there is no causal connection between tha two!

And, you would be right. Right back up in yo muthafuckin ass. Spendin mo’ scrilla on science (please let’s spend more!) aint gonna result up in mo’ suicides. Though tha two is as tightly correlated as any two thangs can be.

[Da above graph is from Tyla Vigen. I aint talkin’ bout chicken n’ gravy biatch yo. His joint – n’ book – Spurious Correlations is straight-up dope naaahhmean, biatch? Yo ass can checkout nuff mo’ correlations, n’ laugh n’ cry n’ laugh n’ cry like a muthafucka. Right back up in yo muthafuckin ass. Start wit tha one bout Nicolas Cage pornos causin playas ta drown!]

Let’s peep another example, cuz it is fresh off tha presses.

All of our asses travel n’ it iz of immense interest ta our asses as ta which airline has high “qualitizzle ratings” when it comes ta performizzle yo. Here is ratings dat came up todizzle…

airline qualitizzle ratings 2015

Most press reports you’ll read bout these performizzle ratings will rap breathlessly bout tha rankin n’ tha movement up or down of a particular airline. What none of dem do is rap bout how tha fuck dis data is calculated. Y’all KNOW dat shit, muthafucka! This type’a shiznit happens all tha time. But, you tha smart-ass Analyst, gonna git yo’ skepticizzle radar up n’ you’ll dig!

Yo crazy-ass startin point can be tha press release from tha source. There is mo’ drill down data there.

Then you’ll poke round ta git into how tha fuck on time arrivals n’ departures is straight-up calculated, whoz ass sets tha standards/formulas, what tha fuck kind of control airlines have when it comes ta settin they schedules, whoz ass decides what tha fuck data gets reported n’ whoz ass audits it, n’ mo’ linez of inquiry before you loot any of all dis bullshit.

When you dig in, you’ll git dat there is straight-up no standard fo’ what tha fuck tha word delay means fo’ realz. Airlines is up in full control of settin tha duration of a gangbangin’ flight. For example, mah flights ta JFK on United routinely arrive “on time,” even though they depart 20 mins late, cuz United addz a half minute ta tha “scheduled flight time.” That paddin then controls what tha fuck is reported (crap basically) fo’ realz. Across airlines there is no standardization of how tha fuck duration between two destinations should be set – so basically tha rankin above is comparin applez n’ monkeys n’ asteroids. Da airlines, like Frontier, dat set tha tightest durations, most accuracy, is bein penalized up in a way fo’ phat behavior.

And, dis is just tha start of tha problem.

Straight-up quickly you’ll realize, tha rankings n’ tha data above is basically garbage. If, all up in tha end of tha dizzle tha purpose of dis data is ta help you cook up a smarta decision, it is miserably failin ta do dat cuz of tha above issues.

Yo, skepticizzle fo’ realz. A phat thang up in a Analyst cuz you know what tha fuck ta use as a source of decisions, n’ what tha fuck not to.

Most data you peep up in tha real ghetto won’t be as obviously wack as you’ll peep up in tha gems shared by Shiznitty Fox Graphics. Da example’s you’ll peep is ghon be mo’ subtle, they is ghon be lookin like they make sense, they will come from sources you trust, from tools you use n’ even implemented yo ass, etc. That’s when you need ta be most vigilant of all ta be a pimped out Analyst.

Here is some steez you can use:

1. Look ta peep if tha conclusion (“insight”) expressed has anythang ta do wit tha data up in front of you, biatch. This, straight-up will only take you a cold-ass lil couple minutes.

2 yo. Here’s a pimped out question: Where did tha data come from, biatch? Tools, countries, people, devices, etc. Known gapz of what’s unknown (particularly relevant up in digital data).

3 fo’ realz. Another one dat you’ll love: What typez of bias might exist up in tha data, biatch? Sample bias, biatch? Samplin bias, biatch? What could cause it ta be incomplete?

4. What principlez you’ve hustled already dat might be fucked up by tha analysis presented, biatch? Correlation/causation is tha one we covered above.

5 fo’ realz. Always, always, always ask dis question: What assumptions was made up in bustin dis analysis?

6. Yo crazy-ass experience. Yo ass gotz a ton of dat shit. Don’t let it git all up in waste.

7. (Added via Slick Rick Hren’s comment below) Dum diddy-dum, here I come biaaatch! Who tha fuck gains or loses from dis analysis, biatch? As wit nuff thangs, follow tha scrilla, power, ballistics.

Numbers 8 all up in 12 was contributed by Ian Frantz…

8. Was tha data collected wit a measurement plan?

9. Was dis data intentionally designed or is it tha bi-thang of another activity?

10 yo. Have we accounted fo’ tha attenuation of data as dat shiznit was collected via a specific medium?

11. Did yo dirty ass create statistically wise boundaries when you chizzle these arbitrary categories?

12. If you handed all tha data, order of operations, software over ta me; could I reproduce tha thangs up in dis biatch?

Numbers 13 all up in 16 was contributed by Rod Jacka…

13. Clarify tha reason fo’ tha claim

14. Test alternatizzle explanations dat can be concluded from tha data or evidence

15. Challenge tha implications n’ consequencez of tha claim. Would it follow dat … will occur cuz of ….

16 fo’ realz. Above all else, question tha question itself 🙂

Numbers 17 all up in 24 was contributed by Jizzy Brazier…

17. If tha data is given as a proportion, what tha fuck was tha raw joints, biatch? (a 50% rise up in shark attacks could be a increase from two ta three cases)

18. If tha data be a raw value, what’s tha proportion, biatch? (a station I pass all up in every last muthafuckin dizzle warns dat 19 escalator incidents have taken place up in tha last year. Shiiit, dis aint no joke. Right back up in yo muthafuckin ass. Since it has round 90 mazillion entrances/exits per year, I be thinkin dis number is pretty irrelevant)

19. Could tha sample done been run multiple times, biatch? (If you git 10 batchez of 20 dem hoes ta test a hair-care product, you’ll likely find dat one of tha batches will produce a high enough satisfaction figure ta advertise)

20. Is tha law of big-ass numbers a problem, biatch? (A UK joint compilez statistics bout response times fo’ Memberz of tha UK Parliament. On average, tha Chronic Jam is tha fastest by a long-ass way; they have 1 MP, compared ta partizzles wit mo’ than 50).

21. Is a proportion penalisin you unfairly, biatch? (Yo crazy-ass site’s treasured 60% conversion rate might be damaged if you was picked up by a major shizzle network n’ given 100,000 low qualitizzle hits, n’ you can put dat on yo’ toast. That don’t make it shitty thang!)

22. Is you bustin a gangbangin’ fair comparison, biatch? (‘Views’ on Facebizzle aren’t tha same as ‘Views’ on YallTube)

23. What aren’t you bein holla’d at, biatch? (Any shizzle rap on some cold-ass lil charitable event dat don’t mention tha volume of fundz raised is likely hidin dat they didn’t make straight-up much)

24. Do tha thug freestylin tha statistic KNOW it, biatch? (Second-hand statistics is often missin crucial caveats)

I’m shizzle there be others. Would you please help me expand dis list by addin steez you’ve hustled ta help brang yo’ healthy skepticizzle forward by addin a cold-ass lil comment below?

When you peep a piece of data, from inside yo’ company or from tha outside, be skeptical up in general. It aint nuthin but tha nick nack patty wack, I still gots tha bigger sack. Well shiiiit, it aint nuthin but a phat trait ta have as a Analyst.

#1B: Skepticizzle should not paralyze you, biatch.

Yo ass is goin ta feel I’m goin ta run all of tha above under a funky-ass bus now, nahmeean, biatch? Please stick wit mah dirty ass.

Da real ghetto aint perfect, n’ yo ass is paid ta help yo’ company (non-profit or for-profit) make smarta decisions every last muthafuckin dizzle (hopefully). One blingin thang at play here, biatch? A decision has ta be made.

Novice Analysts git so caught up in tha skepticizzle dat they become paralyzed cuz if you even lift tha covers under digital analytics a tiny bit, tha way data is collected wit scare tha bejesus outta you, biatch. Oh, n’ offline analytics, biatch? A mazillion times worse fo’ realz. And, tiny data samplez ta boot playa!

Great analysts git phat at one of da most thugged-out critical elementz of our thangs: Timeliness. Da mobilitizzle ta serve up a insight, a specific recommendation, up in a thugged-out duration dat it gonna git a impact on tha bidnizz.

An constipated fuck up is betta than no action at all.

Our thang is ta be skeptical, ta dig n’ KNOW n’ poke n’ prod n’ ta reject tha outrageously wack n’ if it aint outrageously wack then ta git into how tha fuck right it might be all kindsa dat you can cook up a cold-ass lil constipated recommendation.

This post is from 2006: Data Qualitizzle Sucks, Let’s Just Git Over It You’ll learn tha six step process you can use ta overcome tha paralysis.

Here’s a simple way ta be thinkin bout rockin yo’ skepticizzle yo, but still bustin a thugged-out decision.

If you was 100% certain bout tha data, you would immediately recommend ta yo’ company dat they should start makin plans ta build a cold-ass lil colony on tha moon.

If you was 80% certain bout tha data, you could recommend dat they shift tha game wit tha Internationistic Space Station ta start bustin short visits ta tha moon.

If you was 40% certain bout tha data, you could still recommend a triplin of tha investment up in entitizzles on earth dat would study how tha fuck ta live on tha moon.

If you was 20% certain bout tha data, you would go back ta yo’ crew n’ git into what tha fuck strategies you all should put up in place ta git ta at least 40% certainty.

That’s what tha fuck I mean by bein skeptical – yo’ quest is ta git a mo’ concrete feel fo’ where dat certainty lies. Put ya muthafuckin choppers up if ya feel dis! By not bein paralyzed by perfection, I mean bustin a thugged-out decision dat reflects dat certainty cuz tha bidnizz needz timely decision making.

We is on tha topic of bein pimped out Analysts, n’ you can put dat on yo’ toast. Right back up in yo muthafuckin ass. So yo. Here’s a thugged-out detour, n’ yet on theme extension of dat idea.

#2: Da Difference Between Knowledge-Insight-Wisdom.

As a shitload of y’all know, I’m freestylin a twice-a-week short newsletta called Da Marketin – Analytics Intersect. Yo ass can (should!) sign-up fo’ dat shit.

I often find dat most playas whoz ass have tha title Analyst is essentially data collectors n’ data sharers wit most of tha value bein added by dem up in tha process be a tablefied or chartfied summary.

I be a gangsta yo, but y’all knew dat n’ mah newsletta on 29th March, shared a gangbangin’ dunkadelic cartoon dat exquisitely captured tha difference between data – shiznit – knowledge – insight – wisdom. Well shiiiit, it also added a layer, dare I say, wisdom by outlinin tha value of tha thang, salary n’ how tha fuck quickly you can be replaced up in tha thang.

Here’s dat TMAI, up in it’s entirety, I’ll pick tha rap up again n’ again n’ again on tha other side…

TMAI #12

Of all tha cartoons related ta data, n’ analysis, dis one is mah all time favorite…

data shiznit knowledge insight wisdom

[Cartoon by Dizzy Somerville, based on a two pane version by Hugh McLeod.]

Isn’t it incredible, it captures so much bout tha work our phat asses do up in so lil.

I gots a straight-up boner fo’ dis cartoon cuz there be all kindsa nuff insights, :), ta draw from dat shit. Let me focus on one, how tha fuck valuable yo ass is ta yo’ company.

Data: Yo ass be a javascript jock, tha slayer of ETL challenges. Yo ass is tha data hunta n’ gatherer n’ shit. Value: Low. Salary: Low. Replacement: Easy.

Information: Yo ass be a report creator, you fix code up in emergencies. Put ya muthafuckin choppers up if ya feel dis! Value: Low. Salary: Lowish. Replacement: Easy.

Knowledge: Yo ass run a crew of data pukers, you help hook up divisionizzle data needs, yo’ crew merges data sources. Value: Medium. Salary: Medium. Replacement: Takes two months.

Insight: Yo ass hold tha Analyst title, most of tha time you stay tha fuck away from bein peep as a thugged-out data provider, you git invited ta director-level bidnizz meetings. Value: High.Salary: High. Replacement: Hard, six ta nine months.

Wisdom: Yo ass be a Analyst yo, but sit up in a funky-ass bidnizz crew, is yo’ second home, you hook up wit tha CMO every last muthafuckin other week. Value: Priceless. Salary: High times 5. Replacement: Impossible’ish.

Yo, so, what tha fuck thang is you bustin at yo’ company, biatch? Information, biatch? Insight?

Is there mah playas up in yo’ company up in tha analyst crew, or tha marketin crew, whose explicit thang it is ta serve up wisdom?

Yes, you wanna be up in tha Wisdom bidnizz. But, realize how tha fuck hard it is ta do. Yo ass gotta be on a cold-ass lil constant quest fo’ self-improvement, n’ da most thugged-out bangin game you’ll brang ta bear is yo’ bidnizz savvy n’ not data-crunchin prowess. Ironic, no?

Great Analysts solve fo’ Wisdom fo’ realz. And, above n’ beyond what tha fuck you read up in tha newsletter, you can peep how tha fuck both bein skeptical n’ not bein paralyzed helps you git ta Wisdom faster.

One last bonus before we close… If you wanna git a sense fo’ specific salaries, four key chizzlez you gotta make ta git a gangbangin’ fabulous analytics game AND how tha fuck ta git there… Here’s a post you’ll find ta be of value: Web Analytics Game Guide: From Zero To Pimp In Five Steps!

As always, it is yo’ turn now, nahmeean?

Do you smoke we aint skeptical enough bout data floatin round up in our g-units or tha interwebs, biatch? What is strategies you use ta gin n juice yo’ skepticism, biatch? How tha fuck do you torture tha data you see/get, biatch? Is there suttin’ dat works fo’ you particularly well when it comes ta solvin fo’ timeliness, biatch? Is you solvin fo’ Wisdom up in yo’ current thang, biatch? Insights, biatch? What caused yo’ game ta leap from Data ta Hype ta Knowledge faster?

Quit playin’ n’ do what tha fuck I be sayin’! Please share yo’ wisdom, :), experiences, tips, tricks n’ lessons from tha front lines via comments below.

Nuff props, biatch.

Leave a Comment

Your email address will not be published. Required fields are marked *