Statistics Poetry

[PeterFlom]

Avoiding diSASters
by Peter Flom


SAS can be a powerful tool,
But misapply it, and look a fool.
To avoid common errors it's always best
To follow the advice that you'll read next.

SAS has DATA steps and PROCS,
And you can alternate them.
But woe betide the programmer
Who tries to integrate them.

INFORMATS can help you read<br>
Almost any data.
But put them right after the VAR NAME<br>
And do not wait for later.

The semicolon is your friend.
It's how all SAS statements end.
So if you've got some oddball errors
A dot and comma may end your terrors.

A program without comments will run just fine
(SAS really doesn't care)
But six months later will YOU know
Why you put that statement there?

Each variable has got a length
Assigned when it's first mentioned
If you state length explicitly
Your program may be strengthened.

Missing values plague all data
That I have ever seen
But using 999 for them
May distort the value's mean.

Now every MERGE should have its BY
There's a simple reason why
If not, then data sets might get replaced
And you would lose a lot of face.

And finally, please use each PROC
For just what it's intended.
That's all the advice that I have got
So here my poem's ended.

 

[Karabiner]

Meditation on Statistical Method
by J. V. Cunningham

Plato, despair!
We prove by norms
How numbers bear
Empiric forms,

How random wrong
Will average right
If time be long
And error slight,

But in our hearts
Hyperbole
Curves and departs
To infinity.

Error is boundless.
Nor hope nor doubt,
Though both be groundless,
Will average out.

From The Exclusions of a Rhyme: Poems and Epigrams (1960).

 

[Karabiner]

Bayesian versus Frequentist song
by Corrigan Brothers

Toor a loo
toor a loo
toor a loo
toor a listics
this long running spat
in the world of statistics

Now for a long time in the world of stats
Bayesians and Frequentists have a had a spat
Frequentists say probability’s objective
But Bayesians will tell you that its subjective

Toor a loo...

Now the Bayesians are not amateurs
but for them its only random paramaters
while frequentists claim that it has been shown
that paramaters are fixed and unknown

Toor a loo...

Interpretations based on limited frequency
thats was frequentists will have you believe
the old Bayesians think the solution
that over parameters there's probability distribution
Toor a loo...

Written and sung by the Irish band Corrigan Brothers, singers of
There's No One As Irish As Barack Obama.

 

[Karabiner]

Nocione de Estadisica
Statistical Concepts
By Leopoldo de Luis

Note: This is a German's rough translation from Spanish into English.
If you would like to suggest corrections or improvements, then please send me a message.

La estadística es una princesa de azul hielo
que patina por círculos de cálculos metálicos
y arrastra suavemente a sus fríos dominios.

Statistics is a princess of blue ice
that slides by circles of metal calculations
and drags gently to her cold domains.

Entras en sus elipses, en sus cerradas curvas,
en sus circunferencias concéntricas te sumes,
sus ecuaciones ponen espejos a tu imagen,
tus huesos redondeas en sus lentas parábolas,
tus madejas devánanse en su asíntota,
habita en sus incógnitas tu sangre,
eres el leve punto de sus gráficos,
cruzas el seco cielo que se acota
entre su abscisa y su ordenada, eres
el pájaro pequeño que persigue la flecha
de la media aritmética más allá de su nido.

You walk into its ellipses, its closed curves,
in its concentric circumferences you join,
its equations are mirrors to your image,
your bones you round off in its slow parables,
your skeins wind up in its asymptotes,
in its variables dwells your blood,
you're the minor point of its graphics,
you cross the dry sky that is bounded
between its abscissa and its ordinate, you are
the small bird that pursues the arrow
of the arithmetic mean beyond its nest.

Somos el acechado gorrión de la estadística,
caemos en sus redes y el corazón nos tiembla
como trémulas alas, nos sentimos heridos
pero sólo nos cruza un pequeño taladro
para que nos registre la gran computadora.

We are the sparrow waylaid by statistics,
we fall into its nets and our hearts tremble,
shaking wings, we feel wounded,
but we only cross a small hole
so that the great computer registers us.

Hoy me he muerto de hambre, ayer besé tus labios,
mañana seré el preso de un sueño subversivo,
camino con el flanco quemado por un hierro,
soy la res de una inmensa ganadería técnica
y tengo una sonrisa de pena programada.

Today I died of hunger, yesterday I kissed your lips,
tomorrow I will be the prisoner of a subversive dream,
walk with the flank burned by an iron,
I am cattle from a huge technical stock breeding
and I hold a smile of programmed grief.

Yo soy el cero punto y un pequeño guarismo
por ciento de la masa que consume la vida,
apenas si perturbo el nivel de incidencia
y mi amor se regula por un coeficiente.
El cáncer me ha elegido, el infarto me ronda,
la autopista reclama mi cadáver.

I am the zero point and a small
percentage of the mass that consumes life,
hardly disturb the level of incidence,
and my love is regulated by a coefficient.
The cancer has chosen me, the infarction haunts me,
the highway claims my corpse.

Soñé con una pura libertad: bien mirado
pudiera ser su hipótesis, y sé que ya mi muerte
está en una esperanza matemática.

I dreamed with pure freedom: highly regarded
could be its hypothesis, and I know that my death
is an expected value.

http://www.madrimasd.org/cienciaysociedad/poemas/poesia.asp?id=169

 

[Karabiner]

The p Value Song
By Michael Greenacre

Statistics, logistics, cladistics seem to me
To have a common theme scientifically
Economists, biologists with PhD degrees
They all need some proof of their theories
A letter is the key, you’ll see clearly
Not b nor g nor v
But it’s the p!

There’s no values like p values
Like no values I know
Think of something that is not worth proving
An hypothesis that everyone calls Null
If your p is too large to reject it
Then your experiment is rather dull

There’s no values like p values
Especially when they’re low
Don’t be sad if you p’s over .05
Just try again with samples twice the size
Everything is possible, just trust in me
Put your faith in the p

The F test, the z test, the Chi square and the t
And other cryptic terminology
ANOVA, regression, tests distribution-free
They all need some sort of guarantee
So if you find a tiny effect size
The p-value will be a good disguise
There’s no values like p values
The frequentist’s hero
When you get that data modellng feeling
But results you have are not a lot
You will need some stats that are appealing
To show the journal your work is hot

There’s no values like p values
Especially when they are low
Don’t be sad when you p’s over .05
Just try again with samples twice the size
Everything is possible, just trust in me
Put your faith in the p

 

[Karabiner]

Ode to Biostats
By Elizabeth Arend

To the left? To the right? There must be a skew!
Should have paid attention to that math review...
Now I find myself lost in stats,
Clinging to each word spoken in class.

My classmates beside me,
Know just what to do.
But my mere survival is in the hands of
My TI-82.

You just can't imagine the agony and grief
As I struggle to plot another stem, another leaf!
Coffee disappears as I sit at the cafe,
Crunching data sets in back-to-back display.

Stata, I beg you, it just doesn't make sense.
Remind me, which one's a whisker
And which one's a fence?

Logarithms, I feel I live to serve,
But dear Kaplan, dear Meier,
Will I survive the survival curve?

Now the countdown is on,
The mid-term draws near,
And already I'm paralyzed,
Frozen with fear.

Failure is certain from what I can see,
But I'll bet I could calculate it as a probability.

That's right...I think I could.
And know what else I can do?
Marie showed us how to follow Bayes' Rule, too!

It's so cool there's a distribution named after a fish...
And they promised us Nachos,
My favorite dish!

So, think I'm scared of stats?
I'll show you I'm not!
I'll show you with my p values and my QQ plot!

I'll skew those curves to the left and right,
I'll make statistical inferences and stay up all night!
Biostats 620, you ain't got nothing on me.
Just watch me rock those probability.


Biostatistics, Statistical Methods in Public Health I
Department of Biostatistics, Johns Hopkins Bloomberg School of Public Health (2005)
http://www.biostat.jhsph.edu/courses/bio621/misc/poetry.htm

 

[Karabiner]

Confidence Interval - an Epic Poem
by Daan van Schalkwijk

A squire teaches the High King statistics to overcome a crisis of confidence,
at risk to his own life. Just when the High King is at loss in whom to confide, a
squire offers to solve his problems by teaching him about confidence intervals;
but when the King's senior sage protests, he must learn to see what statistics
can and cannot do, or the sage will have him on the gallows.

'My dearest people, oh my people'
The High King sighed in his high hall,
'Oh, how my heart desires to know you,
But could I count you? Not at all!'

'So I have sent for sage and wizard,
To tell me how to serve you best.
But do my confidence they merit?
Oh my poor heart won't give me rest.'

'Oh Sire, Sire' cried a squire,
-Was sternly told to keep his still-,
But he went on, 'Thy need, so dire,
I can alleviate; I will!'

'Then let him speak' declared the High King,
A kindly smile shone in his eyes,
'For the advice this lad will offer,
May chance be canny, if not wise.'

'With your permission, Royal Highness'
Up stepped the squire, with rev'rend bow,
'If it be confidence thy seeking,
I can provide it, I know how.'

'A valiant promise, master squire,
Then let us have it, we're all ears.'
A hidden grin ran through the courtiers,
But our stout squire knew no fears.

'Then well, for starters, take a sample,
Take it as random as one can,
From all the people in thy country,
May they be woman, child, or man.'

'Then from the sample, we must measure,
Whichever size thou longst to know.
It could be tallness, wideness, deepness,
Strength of arm or width of bow.'

And now the trick, your Royal Highness,
That thee will grant thy dream so dear:
It from the averaged sample measure,
Will know your people without fear.'

'I do object, your Royal Highness'
Cried out the kingdom's senior sage,
'What madness does this scoundrel tell us,
So hot of blood and young of age?' ...

(continued at http://scienceandbeyond.blogspot.de/p/confidence-interval-epic-poem.html)

 

[Karabiner]

The Generalized Linear Model
by John K. Kruschke

Straight and proportionate, deep in your core
All is orthogonal, ceiling to floor.
But on the outside the vines creep and twist
'round all the parapets shrouded in mist.

From Chapter 15 of the book Doing Bayesian Data Analysis.
The poem is explained here, using color coding and puppies.

 

[Karabiner]

The Garden’s Forking Paths
by John Schmidt ("with apologies to Robert Frost")

TWO paths diverged on a garden lane,
And knowing I could pick either one
And still seem science-y, saw my gain
And looked down one magnificent vein
To where it twisted toward psych stardom

Then saw the other, not quite as fair,
Though having perhaps a higher claim,
Because it was straighter and lacked in flair;
But then, thought I, who would ev’n care?
I’m sure they’re really about the same…

And both — that morning — equally paid
But wait, it’s just what my grant lacked!
Oh, I’ll take the first for just today
Yet knowing how way leads on to way,
I doubted if I should ever come back.

I shall be telling this with a sigh
In some great conference keynote hence:
Two path diverged in a garden, why — 
I took the one with flair beside,
And that has made all the difference.

The garden of forking paths is a concept proposed by Andrew Gelman and Erik Loken.
If data selection, data preparation and data analysis are contingent on the
dataset at hand, this will lead to biased results, even if the researcher avoids
fishing or p-hacking.

The poem can be found here, the orginal poem for example here.

 

[Karabiner]

Statistics Rap Song (Beyond Average)
by MC SKULE

MC "SKULE" (Spreading Knowledge Using Lyrics & Entertainment)'s debut
Music Video. This song describes, what "Sigma" and "Z-Score" really mean,
and how finding the "Average", while easy, is not enough in statistics.

Do you find yourself asking, What is the point of stats?
Is the time spent learning it worth it just to pass a class?
Well there’s this thing called data, it has characteristics
And making data make sense is the power of statistics
See a long list of numbers, is nice but not quite useful
What we need is to find a way to only focus on what’s crucial
And here’s where you might think; let me just find the average
Well that’s a great start, knowing the mean is an advantage
Compared to just raw data, the mean does us wonders
But relying on just the mean would be quite a blunder

[Chorus] To Be Beyond Average You should go beyond the average
Standard Deviation is something you should want to manage
And calculating Z score might be a chore like cleaning
But I bet you’ll wanna know more when you understand its meaning

You have 2 groups of people, and want to know their ages
So the groups write their ages one by one, on 2 separate pages
Now everyone in group 1, is 20 years old
Which makes their average age 20, lo and behold
Group 2 has a different story, cuz half of the people are 10
But the other half are 30, So the average is 20 again
These 2 groups are quite different, that’s what we want to claim
But just looking at averages, we saw they were the same
Now let’s make a new friend sigma, or standard deviation
It’s a tool that’ll help us make sense of this little complication
It measures how far the average is from each individual
The formula looks scary but applying it is really cool
And the gist of it really is how spread out the numbers are
Which means sigma is higher when all the numbers are really far
So group 2 has a high sigma, and group 1’s is actually zero
And if you can now see why that’s true you’re a sigma super-hero

[Chorus]

Suppose you’ve made a painting, that you’re looking to sell
The mean price is $200, but your painting’s rather swell
So yours will be above the mean, that’s something you can tell
You sold yours for $250, so it seems like you’ve done well $50 above average, so you’re feeling content
But you wonder if your painting was in the top 1 percent
Is 50 above average, really all that high?
Well that depends on whether other prices are nearby
So what we need to quantify is how prices are spread out
But we already have that - sigma without a doubt
So now how many sigmas, above the mean are we?
There’s actually a name for that value: it’s called Z!
So let’s recap what Z score means, the name might seem bizarre
But all it says is how many sigmas from the mean we are
So if sigma is 50, well then that tells us a ton
We’re one sigma above average, so our Z score would be 1
Let’s say you did the math, and found your sigma to be 10
You’re 5 sigmas above average, so the Z is 5 then
The cool thing about Z is that it has its own table
Which lets us solve problems we were previously unable
You’re in the top 1 percent if Z is 2.3 or more
And since our Z is 5 we’re in the 1% for sure
I hope you now can see how to come to this conclusion
And that sigma and Z are no longer a confusion
Don’t get caught up in details, just see the broader vision
Use stats in your life to make better decisions

 

[Karabiner]

Statistics
by William Butler Yeats

'Those Platonists are a curse,' he said,
'God's fire upon the wane,
A diagram hung there instead,
More women born than men.'

Interpretations can for example be found here or here.

 

[Karabiner]

Each Canadian Statistician
By Jeffrey Rosenthal

(to the tune of Glory, Glory, Hallelujah / The Battle Hymn of the Republic / John Brown's Body / Solidarity Forever)

Chorus:
Each Canadian Statistician
Follows in the same tradition
We’re on a very special mission
To treat our data nice!

When we encounter outliers, we throw those points away
But first we stop and smile, and say, "Have a lovely day!"
We've started up a charity to help them find their way
We treat outliers nice!

Equations can be useful, but sometimes they make us freak
It isn't multicultural if every symbol's Greek
We're switching to Cyrilic now, and Hebrew script next week
We treat all symbols nice!

Toutes nos analyses sont traduites tout de suite en français
Nous travaillons très fort pour que personne soit offensé
Ça sera dommage si quelqu'un veut se séparer!
Le français, it is nice!

When data's not symmetric we might take a log or two
But then we say we're sorry if we've inconvenienced you
We tuck in every value with a blanket when we're through
Our transforms must be nice!

We believe in fairness, we treat everyone the same
The world is harsh, and evil inequalities are to blame
From now on, writing equals signs will be our only game
Equalities are nice!

 

[Karabiner]

Written on the Wall of West Forest Temple
by Su Shi

From the side, it's a mountain ridge.
Looking up, it's a single peak.
Far or near, high or low, it never looks the same.
You can't know the true face of Lu Mountain
When you're in its midst.

“This famous Chinese poem from the 11th century suggests that
our perception of reality is limited by our vantage point, which constantly
changes. Likewise, as exemplified here, you should not rely
on a single graphical view to capture the true reality of your data.“
(paraphrased from a blog post written by Minitab Blog Editor)

 

[Karabiner]

Vital Statistics
by Betty L. Levin, Veterans Administration

Births, deaths, and marriages - divorces too,
Veterans in hospitals, to name a few
Are listed in columns, grouped by class -
Statistics gathered from nationwide mass.

With no sign of mirth and no sign of pain
Data are fed the transistorized brain.
Technically, tested, the figures look great.
Charts have been plotted unhampered by faith.

Trends have been studied, analyses printed.
Funds they will cost have yet not been minted.
Statistics are vital - the die is cast.
Our future can only be gleaned from the past.

- But -

If we could survey a vision of life
With all of its joys and all of its strife
Perhaps in our studies, forecasts and such
We'd portray the facts with a human touch.

The American Statistician (1964): Volume 18, Issue 5, page 17

 

[Karabiner]

Ode to Statistics
by Anil Gore and friends

Hail ye, Hail ye.
Hail fellows, happy news
Statistics is for everyone’s use

Daily weather and climate change
Statistics can provide the range
Floods and forest fires too
We can project and alert you

Up and down the roller coaster
Market prices take the swing
Way to manage this fiery dragon
Is with accurate forecasting

Sickness and epidemics all
Snake oilers and claims so tall
Right and wrong need sorting done
Statistics is the only one

Politicians and leaders great
Face periodic voter threat
Surveys and models and computers
Often guess the right answers

What’s life if not a complex puzzle?
Uncertainties and risks so sizzle
For riding the tiger, all ye braves
Numbers guide and statistics saves!

Hail ye, Hail ye. Hail fellows, happy news
Statistics is for everyone’s use

A contribution to the International Year of Statistics 2013.
https://www.worldofstatistics.org/files/2013/04/April-8-2013.pdf

 

[Karabiner]

A hanging histori-o-gram from the Normal Distribution
by David L. Wallace
Celebrating the 250th birthday of the Normal Distribution, 1983.

I
do
sit
at the
very heart
of all of your
theory and art
Pour la statistique
je suis le sel eu le poivre
merci beaucoups a De Moivre
In theory and practice ubiquitous
yet to robustniks, I seem iniquitous
to all my detractors give a coupe de grace
in memory of Marquis Pierre Simon de Laplace
On this celebraton of my demi-semi-millenium
put some of my virtues into a compendium
I have an impressive domain of attraction
a most elegant continued fraction
my negative log density is sweetest
all my cumulants are neatest
I have a straight score
and plenty more
Raise this house
with cheers
for Gauss
and also
many
for
me

The orginal artwork can be found here.
Source: "The Department of Statistics of the University of Chicago celebrates the 275th birthday of the normal distribution" (Nov 12, 2008).

 

[Karabiner]

A Central Limit Poem
by Le Rivenator

From Teaching statistics with poetry (on Twitter).

 

[Karabiner]

Gentlemen, choose your models
by Robert Wherry

Models are fine and statistics are dandy,
But don’t choose too quickly just cause they’re handy
Stick to a model that’s been through the mill
Don’t try something new just for the thrill
A shiny new model is full of allure
But making it work is no sinecure.

The more complex the merrier does not follow
The voluminous output may be hard to swallow
Too many variables and too few cases
Is too much like dueling at ten paces
What’s fit may be error rather than trend
And shrinkage will get you in the end.

Know what you’re doing and do it well
Replictable findings are easy to sell
Be willing to progress one step at the time
A counterfeit dollar’s worth less than a dime
Now that I’ve warned you I’m ready to stop
And let you go back to tending the shop.

Robert J. Wherry Sr. (1975): Underprediction from overfitting: 45 years of shrinkage. Personnel Psychology 28: 1-18
Thanks to Ben Van Cakster for the citation of the poem.

 

[Karabiner]

Extract from en somme, l’inde
by Jacques Darras

…
Les statistiques sont formelles
Une statistique n’est pas un poème
Le poème est une fausse statistique
Les statistiques sont une sale d’attente
Les statistiques attendant qu’on le appelle
Si personne ne lais appelait les statistiques ne bougeraint pas
Les statistiques ont besoin d’un docteur
Attention la poème va se lever
Les statistiques se soignent
Attention le poème se leve...

….
Statistics are formal
A statistic isn’t a poem
The poem is a false statistic
Statistics are a waiting room
Statistics wait for being called
If no one would call them, statistics would not move
Statistics need a doctor
Attention, the poem will get up
Statistics treat themselves
Attention, the poem gets up…

Contained in: Change numérique: Édition intégrale de la revue Change (1968-1983)

Note: This is a German's rough translation from French into English.
If you would like to suggest corrections or improvements, then please send me a message.

 

[Karabiner]

How to determine the cut-off point of your dataset?
by louise_21

| Text | Emotion |
| I like yellow. | Positive |
| I hate blue. | Negative |
| I love yellow. | Positive |
| I dislike orange. | Negative |
| I am ok with any colour. | Neutral |


WORD MEAN
I 0.76
like 0.56
hate 0.45
love 0.03
dislike 0.34
am 0.33
ok 0.22
with 0.10
any 0.02
colour 0.05
yellow 0.20
blue 0.18
orange 0.76
pink 0.05​

From a contribution on talkstats.