back to article Automated, insight cannot be: Jedi master of statistics was good – but beware the daft side

So, farewell then Hans Rosling, educator and "Jedi master of data visualisation". For in a world increasingly addicted to alternative facts, you pioneered software – Gapminder – and a viewer-friendly approach with bubble charts that allowed you to communicate simple important messages about the world through the medium of the …

  1. This post has been deleted by its author

  2. graeme leggett Silver badge

    Rosling and bubbles

    Looking over Rosling's TED presentations, while the bubbles added some extra information but the key information that he was talking about was the X and Y positions. eg education of woman and family size, or life expectancy and number of children.

    The actual bubble size helped to pick China or India out of the other similarly coloured bubbles, but it never detracted from his key points about development of countries.

    1. Anonymous Coward
      Unhappy

      Re: Rosling and bubbles

      I'll go further.

      The article understates Rosling's achievement: not only did he use those bubble-chart presentations to great effect, he and his team developed the software to display them.

      Rosling was a trained statistician, so any implication that he somehow did not know about more rigorous approaches to the presentation of data is wrong. He knew perfectly well how to present data rigorously simply because he must have done so to convince himself and his peers of the validity of his conclusions.

      His talent was then being able to convince non-statisticians (i.e. politicians, aid workers etc) of the same conclusions without putting them off with reams of dry charts.

      On a separate note: what annoys me a lot is that it took more than 10 years for Excel to incorporate even a fraction of the graphing functionality that Rosling pioneered and you still can't replicate one of Rosling's presentations using MS Office. This is a tragedy: all those MS execs watching Rosling's talks and smugly congratulating themselves on how they now 'understand' the third world, but never thinking to themselves how they can support the child at school who wants to become the next Rosling.

      1. Dick Pountain

        Re: Rosling and bubbles

        Absolutely right - Rosling was one of a *very* small group of experts who knew what's going on, cared about it, and was tremendously effective at conveying it to others. I just heard Jeffrey Sachs talk at UCL this morning, and he's another, still with us.

      2. Anonymous Coward
        Anonymous Coward

        Re: Rosling and bubbles

        Why complain about Office? it has probably the most charts in an office suite. You should focus on Qlikview and Tableau, since these new darlings of the reporting work make grand claims for their products when in reality they're behind Excel in terms of charting.

        Without using D3 and extensions, Qlikview is really basic. Better at infographics than anything.

  3. keithpeter Silver badge
    Windows

    Nice article

    Wondering what OA with her industry experience thinks of Edward Tufte's approaches to presentation? I take the points about underlying data quality well.

    https://www.edwardtufte.com/tufte/

    One anecdote: pass rates in exams. Cohort of 100, pass rate 80%, confidence interval approx +- 2sqrt(0.8*0.2/100) which gives 4% either way or an 8% range. So often have I been plunged into emergency action because the *pass rate has dropped* by 2% (i.e. two effing people)! Noise driving management action is *so* much fun.

    (a trend over years would be a whole other thing of course)

  4. Whitter
    Trollface

    "... clear and unequivocal fashion..."

    "Because analysis begins not with visualisations, with the presentation stage, which if it is to mean anything should be about presenting to your audience in clear and unequivocal fashion the results of analysis, but with the data."

    Have we got an icon for irony?

    1. Neil Charles

      Re: "... clear and unequivocal fashion..."

      I was about to quote that exact same paragraph.

      Visual exploration of a dataset is an incredibly powerful tool and as manager of an analytics team, it's where I'd encourage any analyst to start. Whether using commercial software like Tableau*, or free tools like exploratory.io.

      For a quick example of why visual data exploration is so powerful, have a look at Anscombe's quartet.

      https://en.wikipedia.org/wiki/Anscombe's_quartet

      You can hide an awful lot of issues in summary statistics.

      * Actually, not 'like' Tableau. Tableau. It's brilliant for this job.

  5. Whiskers

    Key Performance Indicators?

    You have to choose very carefully what you measure as KPIs. Then make sure you're collecting real data, not stuff 'estimated' at the end of the week by whoever has to fill in the forms.

    KPI = Keeping People Ignorant

  6. Phil O'Sophical Silver badge
    Unhappy

    RIP Hans

    We can argue about how "exact" his presentations were, but let's not lose sight of the fact that he made statics both understandable and fun. One of the few maths presenters I would have gone out of my way to watch. A sad loss.

    1. This post has been deleted by its author

      1. Phil O'Sophical Silver badge

        Re: RIP Hans

        He was never static! ;-)

        Oops, too late to edit the typo. And yes, he was definitely dynamic! :)

    2. Rich 11

      Re: RIP Hans

      One of the few maths presenters I would have gone out of my way to watch.

      Like Johnny Ball, but for adults.

      Come to think of it, both Ball and Rosling could communicate perfectly well to both children and adults without making one group feel left out.

  7. Daggerchild Silver badge

    When is a graph an interface?

    "Graphs are like jokes – if you have to explain them they have failed."

    The old adage applies: Make things as simple as possible, but No Simpler. Sometimes the computer can carry the truth all the way from dataspace to the human realm without spilling any of it, but more complex truths won't survive the journey if the human doesn't try to meet it halfway.

    I like my custom colour-changing fishbone graphs for network traffic state analysis. They're absolute noise to anyone else tho.

  8. Schultz
    Alert

    Pie charts ...?

    The serious presenter long ago moved to sphere charts -- why else did we pay for those 3D TVs?

    No, seriously, google it!

  9. druck Silver badge
    Thumb Down

    Hockey sticks

    Graphing tip for climate researchers (I'm looking at you NASA) if you want to have any credibility; don't release updated graphs where you've just rotated the data around the axis to make the past colder and the present warmer. https://www.theregister.co.uk/2008/05/02/a_tale_of_two_thermometers/

  10. John Smith 19 Gold badge
    Unhappy

    "Do not, however, mistake good communication skills for real analysis:"

    This point right here.

    "Presentation skills" are the way of the dark side Marketing

    Analysis is the way of real management.

    Farewell Hans. You had points to make and you made them effectively. Let's see if anyone paid attention.

  11. Anonymous Coward
    Anonymous Coward

    Restrictions and Measure vs Dimension

    The thing that always bothered me about the wealth vs life expectancy presentation is that is never once highlights the impacts of the very strong influence of Infant Mortality Rate and whether or not a lower bound is included (i.e. of people who survive past infant mortality cut-off age...THIS is the avg. life expectancy)...of course this could be highlighted using ranges and medians as well

    IMHO time-dependent measures should be primarily considered as a dimension/segmentation as opposed to a KPI. For instances, one of my clients wanted to know "avg age, in months, that a customer has been with registered with us" which is obviously always INCREASING if the bulk of acquisitions are historic or DECREASING if bulk are happening in more recent periods...very difficult for people to get their heads around that so better to segment as "New", "1-3 Months",...,"12 Months+", for instance. Especially if the chart itself has a time axis. (yes, yes in a proportional sense the same issue arises but at least you can SEE that influence and for non-proportional KPI's, gives far better insight)

    Great presentation and obviously a man with great and respectable/enviable passion but that twist ending of "but here's what you should consider of the underlying data" would have been nice.

    1. Cynic_999

      Re: Restrictions and Measure vs Dimension

      "Wealth" is also difficult to portray. An annual income of £10000 will make you pretty wealthy in a country where a good meals costs 50p and a decent size house can be had for £5000. Perhaps instead of showing as a monetary value, it should be depicted as the equivalent income in terms of pints of beer.

  12. allthecoolshortnamesweretaken

    "As for axes, these are a perennial source of problems. Show point of origin on both x and y scales, and you risk turning all but the most egregious variations (in say the FTSE 100, for instance) as a near straight line. But truncate them – plot the FTSE over a range from 6800 to 7200, for instance – and you appear to show major fluctuations where only minor ones exist."

    And this, boys and girls, is how "chart analysis" works.

  13. rusty_wheat

    I don't know how much statistical training Hans had but I know he was a physician and professor at the Karolinksa institute. I don't think he had an advanced degree in statistics.

  14. TURN ON TUNE IN DROP TABLES

    "Graphs are like jokes – if you have to explain them they have failed."

    Jokes about graph theory tend to fail.

  15. Anonymous Coward
    Anonymous Coward

    "when a manager calls up a chart with no familiarity with the underlying data"

    you enter the world of 'vanity metrics' where anything can happen as long as the manager thinks they are getting enough shiny graphics to build their fiefdom while making your life a miserable data collecting frenzy instead of actually working...

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon