Tufte’s Fundamental Principles of Analytical Design

Carte Figurative des pertes successives en hommes de l’Armée Française dans la campagne de Russie 1812–1813, (M.Minard, 1869)

In his 2006 offer­ing, Edward Tufte high­lights his Fun­da­men­tal Prin­ci­ples of Ana­lyt­i­cal Design. In this arti­cle, I will show­case these prin­ci­ples and pro­vide an illus­trat­ing exam­ple. (The con­trast and com­par­i­son with the exam­ples pro­vid­ed in Beau­ti­ful Evi­dence show­case the pos­si­ble vari­a­tions in inter­ests and aes­thet­ics).

Through­out, Tufte’s com­ments and insights are shown in block quotes; our own com­ments appear as reg­u­lar text. When­ev­er pos­si­ble, exam­ples are sourced and linked to exter­nal sources, which may pro­vide more con­text and detailed infor­ma­tion.

Fundamental Principles of Analytical Design

Why do we dis­play evi­dence in a report, in a news­pa­per arti­cle, online? What’s the fun­da­men­tal rea­son or goal of our charts and graphs? Tufte sug­gests that we present evi­dence to assist our think­ing process­es (p. 137). In this regard, his prin­ci­ples are uni­ver­sal – a strong argu­ment can be made that they are depen­dent nei­ther on tech­nol­o­gy nor cul­ture. Rea­son­ing and com­mu­ni­cat­ing our thoughts are inter­twined with our lives in a causal and dynam­ic mul­ti­vari­ate Uni­verse;1 what­ev­er cog­ni­tive skills allow us to live and evolve can also be brought to bear on the pre­sen­ta­tion of evi­dence.

Tufte also high­lights a par­tic­u­lar sym­me­try to visu­al dis­plays of evi­dence, that con­sumers should be seek­ing exact­ly what pro­duc­ers should be pro­vid­ing, name­ly

  • mean­ing­ful com­par­isons;
  • causal net­works and under­ly­ing struc­ture;
  • mul­ti­vari­ate links;
  • inte­grat­ed and rel­e­vant data;
  • hon­est doc­u­men­ta­tion, and
  • pri­ma­ry focus on con­tent.

Phys­i­cal sci­ence dis­plays tend to be less descrip­tive and ver­bal, more visu­al and quan­ti­ta­tive; these trends tend to be reversed when deal­ing with evi­dence dis­plays about human behav­iour;2 in spite of this, Tufte argues that his prin­ci­ples of ana­lyt­i­cal design can also be applied to social sci­ence and med­i­cine. To demon­strate the uni­ver­sal­i­ty of his prin­ci­ples, Tufte describes in detail how they are applied in a visu­al dis­play by Charles Joseph Minard (see fig­ure below).

Carte Figurative des pertes successives en hommes de l'Armée Française dans la campagne de Russie 1812-1813, (M.Minard, 1869).
Carte Fig­u­ra­tive des pertes suc­ces­sives en hommes de l’Armée Française dans la cam­pagne de Russie 1812–1813, (M.Minard, 1869).

His lengthy analy­sis of the image is well worth the read (pp. 122–139) – it will not be repeat­ed here (I must con­fess that the chart leaves me some­what … unex­cit­ed).

Rather, I will illus­trate the prin­ci­ples with the help of the fol­low­ing image from the Gap­min­der Foun­da­tion.

Life expectancy and income per capita in 2012, by nation (Gapminder Foundation).
Life expectan­cy and income per capi­ta in 2012, by nation (Gap­min­der Foun­da­tion).

It is a bub­ble chart that plots the 2012 life expectan­cy, adjust­ed income per per­son in USD (log-scaled), pop­u­la­tion, and con­ti­nen­tal mem­ber­ship for 193 UN mem­bers and 5 oth­er coun­tries, using the lat­est avail­able data (2011). A high-res­o­lu­tion ver­sion of the image can be found on the Gap­min­der web­site.


First Prin­ci­ple
Show com­par­isons, con­trasts, dif­fer­ences. (p. 127)

The Fun­da­men­tal ana­lyt­i­cal act in sta­tis­ti­cal rea­son­ing is to answer the ques­tion “Com­pared with what?” Whether we are eval­u­at­ing changes over space or time, search­ing big data bases, adjust­ing and con­trol­ling for vari­ables, design­ing exper­i­ments, spec­i­fy­ing mul­ti­ple regres­sions, or doing just about any kind of evi­dence-based rea­son­ing, the essen­tial point is to make intel­li­gent and appro­pri­ate com­par­isons [empha­sis added]. Thus, visu­al dis­plays […] should show com­par­isons. (p. 127)

Com­par­isons come in var­ied flavours: for instance, one could com­pare a

  • unit at a giv­en time against the same unit at a lat­er time;
  • unit’s com­po­nent against anoth­er of its com­po­nents;
  • unit against anoth­er unit,

or any num­ber of com­bi­na­tions of these flavours. Not every com­par­i­son will turn out to be insight­ful, but avoid­ing com­par­isons alto­geth­er is equiv­a­lent to pro­duc­ing dis­plays built from a sin­gle datum, and… well, what’s the point, then?

    Health and Wealth of Nations

    Where to begin? First, note that each bub­ble rep­re­sents a dif­fer­ent coun­try, and that the loca­tion of each bubble’s cen­tre is a pre­cise point cor­re­spond­ing to the country’s life expectan­cy and its GDP per capi­ta. The size of the bub­ble is cor­re­lat­ed with the country’s pop­u­la­tion, while its colour is linked to con­ti­nen­tal mem­ber­ship.

    The chart’s com­pass pro­vides a handy tool for com­par­i­son:

    • a bub­ble to the right (resp. the left) rep­re­sents a wealth­i­er (resp. poor­er) coun­try;
    • a bub­ble above (resp. below) rep­re­sents a health­i­er (resp. sick­er) coun­try.

    For instance, a com­par­i­son between Japan, Ger­many and the USA shows that Japan is health­i­er than Ger­many, which is itself health­i­er than the USA (as deter­mined by life expectan­cy), while the USA are wealth­i­er than Ger­many, which is itself wealth­i­er than Japan (as deter­mined by GDP per capi­ta) (see below).

    Comparison between Japan, Germany, and the USA, Health and Wealth of Nations, 2012.
    Com­par­i­son between Japan, Ger­many, and the USA, Health and Wealth of Nations, 2012.

    While there is no rea­son to expect that it should be the case, it is nev­er­the­less pos­si­ble for two coun­tries to have rough­ly the same health and the same wealth: con­sid­er Indone­sia and Fiji, or India and Tuvalu, for instance (see below). In each pair, the cen­tres of both bub­bles over­lap: any dif­fer­ence in the data must be found in the bub­bles’ area or their colour.

    Comparison between Indonesia and Fiji, and India and Tuvalu, Health and Wealth of Nations, 2012.
    Com­par­i­son between Indone­sia and Fiji, and India and Tuvalu, Health and Wealth of Nations, 2012.

    Coun­tries can also be com­pared against world val­ues for life expectan­cy and GDP per capi­ta (the com­par­isons for aver­age con­ti­nen­tal mem­ber­ship or pop­u­la­tion are less obvi­ous­ly mean­ing­ful, in this case). The world’s mean life expectan­cy and income per per­son are traced in light blue. Wealth­i­er, health­i­er, poor­er, and sick­er are rel­a­tive terms, but we can also use them to clas­si­fy the world’s nations with respect to these mean val­ues.3

    Comparison of data points against life expectancy of the world, Health and Wealth of Nations, 2012.
    Com­par­i­son of data points against life expectan­cy of the world, Health and Wealth of Nations, 2012.

    Comparison of data points against GDP per capita of the world, Health and Wealth of Nations, 2012.
    Com­par­i­son of data points against GDP per capi­ta of the world, Health and Wealth of Nations, 2012.

Causality, Mechanism, Structure, Explanation

Sec­ond Prin­ci­ple
Show causal­i­ty, mech­a­nism, expla­na­tion, sys­tem­at­ic struc­ture. (p. 128)

Yet often the rea­son that we exam­ine evi­dence is to under­stand causal­i­ty, mech­a­nism, dynam­ics, process, or sys­tem­at­ic struc­ture [empha­sis added]. Sci­en­tif­ic research involves causal think­ing, for Nature’s laws are causal laws. […] Rea­son­ing about reforms and mak­ing deci­sions also demands causal log­ic. To pro­duce the desired effects, we need to know about and gov­ern the caus­es; thus “pol­i­cy-think­ing is and must be causal­i­ty-think­ing”.4 (p. 128)

Sim­ply col­lect­ing data may pro­voke thoughts about cause and effect: mea­sure­ments are inher­ent­ly com­par­a­tive, and com­par­isons prompt­ly lead to rea­son­ing about var­i­ous sources of dif­fer­ences and vari­abil­i­ty. (p. 128)

In essence, this is the core prin­ci­ple behind data visu­al­iza­tion: the dis­play needs to explain some­thing, it needs to pro­vide links between cause and effect, it needs to tell a mean­ing­ful sto­ry.

If the visu­al­iza­tion could be removed at a lat­er stage with­out chang­ing the under­ly­ing mes­sage, then that chart should not have been includ­ed in the first place, no mat­ter how pret­ty and mod­ern it looks, nor how cost­ly it was to pro­duce.

    Health and Wealth of Nations

    At a quick glance, the rela­tion between the log of the income per per­son and life expectan­cy seems to be increas­ing rough­ly lin­ear­ly. The exact para­me­ter val­ues are not known (and I can­not esti­mate them ana­lyt­i­cal­ly as I do not have access to the data), but an approx­i­mate line-of-best-fit has been added to the fig­ure below. Charts with this form have been found in oth­er dis­ci­plines before.

    Linear relationship in the Health and Wealth of Nations, 2012.
    Lin­ear rela­tion­ship in the Health and Wealth of Nations, 2012.

    The four quad­rants cre­at­ed by the world’s life expectan­cy and its GDP per capi­ta are not all cre­at­ed equal: naive­ly, we might have expect­ed that each of the quad­rants would con­tain about 25% of the world’s coun­tries (although the large pop­u­la­tion size of giants like Chi­na and India mud­dle the pic­ture some­what), how­ev­er, there is one quad­rant which is sub­stan­tial­ly under-rep­re­sent­ed (see below). Is it sur­pris­ing that there should be so few “wealth­i­er” and “sick­er” coun­tries? Could it be argued that Rus­sia and Kaza­khstan are too near to the sep­a­ra­tors to real­ly be con­sid­ered clear-cut mem­bers of the quad­rant?

    The wealthier and sicker quadrant, Health and Wealth of Nations, 2012.
    The wealth­i­er and sick­er quad­rant, Health and Wealth of Nations, 2012.

    In the same vein, when we con­sid­er the data visu­al­iza­tion as a whole, there seems to be one group of out­liers below the main trend, to the right (and pos­si­bly one group above the main trend, to the left) which cries out for an expla­na­tion: South Africa, for instance, has a rel­a­tive­ly high GDP per capi­ta but a low life expectan­cy (poten­tial­ly, income dis­par­i­ty between a poor major­i­ty and a sub­stan­tial­ly rich­er minor­i­ty might help push the dot to the right, while the low­er life expectan­cy of the major­i­ty dri­ves the over­all life expectan­cy to the bot­tom). Could wars, famines, reces­sions, and epi­demics be recov­ered from the move­ment of the bub­bles over mul­ti­ple years?

    Outliers in the Health and Wealth of Nations, 2012.
    Out­liers in the Health and Wealth of Nations, 2012.

Multivariate Analysis

Third Prin­ci­ple
Show mul­ti­vari­ate data; that is, show more than 1 or 2 vari­ables. (p. 130)

Near­ly all the inter­est­ing worlds (phys­i­cal, bio­log­i­cal, imag­i­nary, human) we seek to under­stand are inevitably mul­ti­vari­ate in nature. (p. 129)

The analy­sis of cause and effect, ini­tial­ly bivari­ate, quick­ly becomes mul­ti­vari­ate through such nec­es­sary elab­o­ra­tions as the con­di­tions under which the causal rela­tion holds, inter­ac­tion effects, mul­ti­ple caus­es, mul­ti­ple effects, causal sequences, sources of bias, spu­ri­ous cor­re­la­tion, sources of mea­sure­ment error, com­pet­ing vari­ables, and whether the alleged cause is mere­ly a proxy or a mark­er vari­able.5 p. 129)

Rea­son­ing about evi­dence should not be stuck in 2 dimen­sions, for the world we seek to under­stand is pro­found­ly mul­ti­vari­ate [empha­sis added]. (p. 130)

Alert read­ers may ques­tion the ulti­mate valid­i­ty of this prin­ci­ple: after all, doesn’t Occam’s Razor warn us that “it is futile to do with more things that which can be done with few­er”?6 Seen in the right light, that seems like a fair­ly strong admo­ni­tion to stay away from mul­ti­vari­ate analy­sis.

This inter­pre­ta­tion depends, of course, on what it means to “do with few­er”: are we attempt­ing to “do with few­er”, or to “do with few­er”?. If it’s the for­mer, then we can pro­duce any num­ber of uni­vari­ate and bivari­ate charts to rep­re­sent the data (which in itself starts to bor­der on a mul­ti­vari­ate dis­play, although the num­ber of such charts can bal­loon quite quick­ly), but any sig­nif­i­cant link between 3 and more vari­ables is unlike­ly to be shown, which dras­ti­cal­ly reduces the explana­to­ry pow­er of the charts. If it’s the lat­ter, the dif­fi­cul­ty evap­o­rates: we sim­ply retain as much fea­tures as nec­es­sary to main­tain the desired explana­to­ry pow­er.

    Health and Wealth of Nations

    Only 4 vari­ables are rep­re­sent­ed in the dis­play, which we could argue just bare­ly qual­i­fies the data as mul­ti­vari­ate. The pop­u­la­tion size seems uncor­re­lat­ed with both of the axes’ vari­ates, unlike con­ti­nen­tal mem­ber­ship: there is a clear divide between the West, most of Asia, and Africa (see below). This “clus­ter­ing” of the world’s nations cer­tain­ly fits with com­mon wis­dom about the state of the plan­et, which pro­vides some lev­el of val­i­da­tion for the dis­play.

    Clusters in the Health and Wealth of Nations, 2012.
    Clus­ters in the Health and Wealth of Nations, 2012.

    Oth­er vari­ables could also be con­sid­ered or added, notably the year, allow­ing for bub­ble move­ment: one would expect that life expectan­cy and GDP per capi­ta have both been increas­ing over time. The Gap­min­der Foundation’s online tool can build bub­ble charts with oth­er vari­ates, lead­ing to inter­est­ing infer­ences and con­clu­sions.

Integration of Evidence

Fourth Prin­ci­ple
Com­plete­ly inte­grate words, num­bers, images, dia­grams. (p. 131)

The evi­dence doesn’t care what it is – whether word, num­ber, image. In rea­son­ing about sub­stan­tive prob­lems, what mat­ters entire­ly is the evi­dence, not par­tic­u­lar modes of evi­dence [empha­sis added]. (p. 130)

Words, num­bers, pic­tures, dia­grams, graph­ics, charts, tables belong togeth­er [empha­sis added]. Excel­lent maps, which are the heart and soul of good prac­tices in ana­lyt­i­cal graph­ics, rou­tine­ly inte­grate words, num­bers, line-art, grids, mea­sure­ment scales. (p. 131)

Tables of data might be thought of as para­graphs of num­bers, tight­ly inte­grat­ed with the text for con­ve­nience of read­ing rather than seg­re­gat­ed at the back of a report. […] Images and tables used in pub­lic pre­sen­ta­tions should be anno­tat­ed with words explain­ing what is going on. In explorato­ry data analy­sis, how­ev­er, the inte­gra­tion of data needs to be thought through. Per­haps the num­ber of data points may stand alone for a while, so we can get a clean look at the data, although tech­niques of lay­er­ing and sep­a­ra­tion may simul­ta­ne­ous­ly allow a clean look as well as bring­ing oth­er infor­ma­tion into the scene. (p. 131)

When authors and researchers select a sin­gle spe­cif­ic method or mode of infor­ma­tion dur­ing the inquiries, the focus switch­es from “can we explain what’s hap­pen­ing?” to “can the method we select­ed explain what’s hap­pen­ing?”. There is an art to select­ing meth­ods, and expe­ri­ence and exper­tise can often sug­gest rel­e­vant meth­ods, but “when all one has is a ham­mer, every­thing looks like a nail”, as the say­ing goes: the goal should be to use what­ev­er (and all) evi­dence is nec­es­sary to explain “what’s hap­pen­ing”. If that goal is met, it makes no dif­fer­ence which modes of evi­dence were used.

    Health and Wealth of Nations

    The var­i­ous details attached to the chart (such as coun­try names, font sizes, axes scale, grid, and world land­marks) pro­vide sub­stan­tial ben­e­fits when it comes to con­sum­ing the dis­play. They may become lost in the back­ground, with the effect that they are tak­en for grant­ed. Com­pare the dis­play obtained from (near­ly) the same data, but with­out inte­gra­tion of evi­dence (see below).

    Health and Wealth of Nations, 2012, without integration of evidence.
    Health and Wealth of Nations, 2012, with­out inte­gra­tion of evi­dence.

    Not near­ly as com­pelling, eh? What’s miss­ing?


Fifth Prin­ci­ple
Thor­ough­ly describe the evi­dence. Pro­vide a detailed title, indi­cate the authors and spon­sors, doc­u­ment the data sources, show com­plete mea­sure­ment scales, point out rel­e­vant issues. (p. 133)

The cred­i­bil­i­ty of an evi­dence pre­sen­ta­tion depends sig­nif­i­cant­ly on the qual­i­ty and integri­ty of the authors and their data sources. Doc­u­men­ta­tion is an essen­tial mech­a­nism of qual­i­ty con­trol for dis­plays of evi­dence. Thus authors must be named, spon­sors revealed, their inter­ests and agen­da unveiled, sources described, scales labeled, details enu­mer­at­ed [empha­sis added]. (p. 132)

Depend­ing on the con­text, ques­tions and items to address could include:

  • What is the title/subject of the visu­al­iza­tion?
  • Who did the analy­sis?
  • Who cre­at­ed the visu­al­iza­tion? (if dis­tinct from analyst(s))
  • When was the visu­al­iza­tion pub­lished?
  • Which ver­sion of the visu­al­iza­tion is ren­dered here?
  • Where did the under­ly­ing data come from?
  • Who spon­sored the dis­play?
  • What assump­tions were made dur­ing data pro­cess­ing and clean-up?
  • What colour schemes, leg­ends, scales are in use in the chart?

It’s not obvi­ous whether all this infor­ma­tion can fit inside a sin­gle chart in some cas­es. But, keep­ing in mind the Prin­ci­ple of Inte­gra­tion of Evi­dence, charts should not be pre­sent­ed in iso­la­tion in the first place, and some of the rel­e­vant infor­ma­tion can be pro­vid­ed in the text, on the web­page, or in an accom­pa­ny­ing doc­u­ment. This is espe­cial­ly impor­tant when it comes to dis­cussing the method­olog­i­cal assump­tions used for data col­lec­tion, pro­cess­ing, and analy­sis. An hon­est assess­ment may require siz­able amounts of text, and it may not be rea­son­able to include that infor­ma­tion with the dis­play (in that case, a link to the accom­pa­ny­ing doc­u­men­ta­tion should be pro­vid­ed).

Pub­licly attrib­uted author­ship indi­cates to read­ers that some­one is tak­ing respon­si­bil­i­ty for the analy­sis; con­verse­ly, the absence of names sig­nals an eva­sion of respon­si­bil­i­ty. […] Peo­ple do things, not agen­cies, bureaus, depart­ments, divi­sions [empha­sis added]. (pp. 132–133)

    Health and Wealth of Nations
    • What is the title/subject of the visu­al­iza­tion?
    • The health and wealth of nations in 2012, using the lat­est avail­able data (2011).

    • Who did the analy­sis? Who spon­sored the dis­play? Who cre­at­ed the visu­al­iza­tion?
    • The analy­sis was done by the Gap­min­der Foun­da­tion; the map lay­out was cre­at­ed by Paulo Fau­sone. No data regard­ing the spon­sor is found on the chart or in the doc­u­men­ta­tion. The rel­e­vant Wikipedia arti­cle states that “the Gap­min­der Foun­da­tion is a non-prof­it ven­ture reg­is­tered in Stock­holm, Swe­den, that pro­motes sus­tain­able glob­al devel­op­ment and achieve­ment of the Unit­ed Nations Mil­len­ni­um Devel­op­ment Goals by increased use and under­stand­ing of sta­tis­tics and oth­er infor­ma­tion about social, eco­nom­ic and envi­ron­men­tal devel­op­ment at local, nation­al and glob­al lev­els.” It seems plau­si­ble that there is no exter­nal spon­sor, but that is no cer­tain­ty.

    • When was the visu­al­iza­tion pub­lished? Which ver­sion of the visu­al­iza­tion is ren­dered here?
    • The 11th ver­sion of this chart was pub­lished in Sep­tem­ber 2012. It is the lat­est avail­able ver­sion as of Octo­ber 2016.

    • Where did the under­ly­ing data come from? What assump­tions were made dur­ing data pro­cess­ing and clean-up?
    • Typ­i­cal­ly, the work that goes into prepar­ing the data is swept under the car­pet in favour of the visu­al­iza­tion itself; there are no explic­it source of data on this chart, for instance. How­ev­er, there is a URL in the leg­end box that leads to detailed infor­ma­tion.
      gap42For most coun­tries, life expectan­cy data was col­lect­ed from the Human Mor­tal­i­ty data­base, the UN Pop­u­la­tion Divi­sion World Pop­u­la­tion Prospects, files from his­to­ri­an James C. Riley, the Human Life Table data­base, data from diverse nation­al sta­tis­ti­cal agen­cies, the CIA World Fact book, the World Bank, and the South Sudan Nation­al Bureau of Sta­tis­tics.

      Bench­mark 2005 GDP data was derived via regres­sion analy­sis from Inter­na­tion­al Com­par­i­son Pro­gram data for 144 coun­tries, and extend­ed to oth­er juris­dic­tions using anoth­er regres­sion against data from the UN Sta­tis­ti­cal Divi­sion, Mad­di­son Online, the CIA World Fact book, and esti­mates from the World Bank. The 2012 val­ues were then derived from the 2005 bench­marks using long-term growth rates esti­mate from Mad­di­son Online, Bar­ro & Ursua, the Unit­ed Nations Sta­tis­ti­cal Divi­sion, the Penn World Table (mark 6.2), the Inter­na­tion­al Mon­e­tary Fund’s World Eco­nom­ic Out­look data­base, the World Devel­op­ment Indi­ca­tors, Euro­stat, and nation­al sta­tis­ti­cal offices or some oth­er spe­cif­ic pub­li­ca­tions.
      Pop­u­la­tion esti­mates were col­lat­ed from the Unit­ed Nations Pop­u­la­tion Divi­sion World Pop­u­la­tion Prospects, Mad­di­son Online, Mitchell’s Inter­na­tion­al His­tor­i­cal Sta­tis­tics, the Unit­ed Nations Sta­tis­ti­cal Divi­sion, the US Cen­sus Bureau, nation­al sources, undoc­u­ment­ed sources, and “guessti­mates”. Exact fig­ures for coun­tries with a pop­u­la­tion below 3 mil­lion inhab­i­tants were not need­ed as this marked the low­er end of the chart res­o­lu­tion.

    • What colour schemes, leg­ends, scales are in use in the chart? 
    • The Leg­end Inset is fair­ly com­pre­hen­sive:

      Legend, Health and Wealth of Nations, 2012.
      Leg­end, Health and Wealth of Nations, 2012.

      Per­haps the last item of note is that the scale of the axes dif­fers: life expectan­cy is mea­sured lin­ear­ly, where­as GDP per capi­ta is mea­sured on a log­a­rith­mic scale. 

    Content Counts Most of All

    Sixth Prin­ci­ple
    Ana­lyt­i­cal pre­sen­ta­tions ulti­mate­ly stand of fall depend­ing on the qual­i­ty, rel­e­vance, and integri­ty of their con­tent. (p. 136)

    The most effec­tive way to improve a pre­sen­ta­tion is to get bet­ter con­tent [empha­sis added] […] design devices and gim­micks can­not sal­vage failed con­tent. (p. 136)

    The first ques­tions in con­struct­ing ana­lyt­i­cal dis­plays are not “How can this pre­sen­ta­tion use the col­or pur­ple?” Not “How large must the logo­type be?” Not “How can the pre­sen­ta­tion use the Inter­ac­tive Vir­tu­al Cyber­space Pro­to­col Dis­play Tech­nol­o­gy?” Not dec­o­ra­tion, not pro­duc­tion tech­nol­o­gy. The first ques­tion is “What are the con­tent-rea­son­ing tasks that this dis­play is sup­posed to help with?” (p. 136)

    A com­pelling nar­ra­tive, which may not be the one that was ini­tial­ly expect­ed to emerge from a sol­id analy­sis of sound data, is the name of the game. Sim­ply speak­ing, the visu­al dis­play should assist in explain­ing the sit­u­a­tion at hand and at answer­ing the orig­i­nal ques­tions.

      Health and Wealth of Nations

      How would we answer the fol­low­ing ques­tions:

      • Do we observe sim­i­lar pat­terns every year?
      • Does the shape of the rela­tion­ship between life expectan­cy and log-GDP per capi­ta vary con­tin­u­ous­ly over time?
      • Do coun­tries ever migrate large dis­tances in the dis­play over short peri­ods?
      • Do excep­tion­al events affect all coun­tries sim­i­lar­ly?
      • What are the effects of seces­sion or annex­a­tion?

      The 2012 Health and Wealth of Nations data rep­re­sent a sin­gle datum in the gen­er­al space of data visu­al­iza­tions; bet­ter con­tent means get­ting data for more than just 2012.

      Health and Wealth of Nations, 2013.
      Health and Wealth of Nations, 2013.

About Dr. Idlewyld 7 Articles
As a youth, Dr. Idlewyld used to read everything he could lay his hands on and he was in a band. For years, he believed that the NHL would have come calling if he hadn't broken his leg as a kid in a hilarious skiing mishap. Nowadays, whatever's left of his hair is slowly turning grey, and that can only mean one thing: he's had the chance to work on plenty of quantitative projects, providing expertise in operations research methods, data science and predictive analytics, stochastic and statistical modeling, and simulations. So he's got that going for him, which is nice. He's not keen on buzzwords, but overall he's glad to see interest in analytical endeavours grow. In the final analysis, he thinks that insights and discoveries are within everyone's reach, and that he would have made a great goalie.