Humanist Discussion Group

Humanist Archives: July 2, 2021, 8:01 a.m. Humanist 35.120 - phantoms of Big Data

				                  Humanist Discussion Group, Vol. 35, No. 120.
        Department of Digital Humanities, University of Cologne
                   		Hosted by DH-Cologne
                       www.dhhumanist.org
                Submit to: humanist@dhhumanist.org


    [1]    From: scholar-at-large@bell.net 
           Subject: Apophenia >>> Re: [Humanist] 35.118: phantoms of Big Data (85)

    [2]    From: Henry Schaffer 
           Subject: Re: [Humanist] 35.118: phantoms of Big Data (12)


--[1]------------------------------------------------------------------------
        Date: 2021-07-01 15:05:57+00:00
        From: scholar-at-large@bell.net 
        Subject: Apophenia >>> Re: [Humanist] 35.118: phantoms of Big Data

Willard,

Your posting of excerpts from the review of “the phantom pattern problem”
reminds me of a passage in Geoffrey Rockwell and Stéfan Sinclair’s
_Hermeneutitica: Computer-Assisted Interpretation in the Humanities_ (2016).


With enough data one can get spurious correlations, as there is always something
that has the same statistical profile as the phenomenon you are studying. This
is the machine equivalent to apophenia, the human tendency to see patterns
everywhere, which is akin to what Umberto Eco explores in _Interpretation and
Overinterpretation_ (1992).


The learning may have to be continuous…


~  ~  ~  ~  ~  ~  ~
François Lachance
Scholar-at-large
Wannabe Professor of Theoretical and Applied Rhetoric
http://www.chass.utoronto.ca/~lachance
https://berneval.hcommons.org

to think is often to sort, to store and to shuffle: humble, embodied tasks






> On Jul 1, 2021, at 1:42 AM, Humanist  wrote:
>
>                  Humanist Discussion Group, Vol. 35, No. 118.
>        Department of Digital Humanities, University of Cologne
>                               Hosted by DH-Cologne
>                       www.dhhumanist.org
>                Submit to: humanist@dhhumanist.org
>
>
>
>
>        Date: 2021-07-01 05:34:10+00:00
>        From: Willard McCarty 
>        Subject: phantoms of Big Data
>
> This is to draw your attention (if you wish it drawn) to a review of the
> following book in Metascience 30, 335–338 (2021), Frank Cabrera,
> "Correlation isn’t good enough: causal explanation and Big Data", on
> Gary Smith and Jay Cordes, The phantom pattern problem: The mirage of
> Big Data (New York: Oxford University Press, 2020). From the review:
>
>> In their book The Phantom Pattern Problem: The Mirage of Big Data,
>> economist Gary Smith and statistician Jay Cordes demonstrate with an
>> arsenal of examples derived from such diverse areas as sports,
>> finance, gambling, astronomy, medicine, etc., the pervasiveness of
>> so-called phantom patterns, i.e., coincidental past correlations that
>> have little to no future predictive value. In addition to
>> illustrating the ease with which humans can be fooled by these
>> coincidental correlations, the authors provide many strategies
>> throughout the book to avoid being misled by phantom patterns. The
>> book is written in a highly engaging, conversational style, is
>> largely non-technical, and is therefore suitable for a wide range of
>> audiences.
>
> The reviewer cites from the book the example of "an observed correlation
> between the rise in reported murders and the rise in iPod sales between
> 2004 and 2006". Cabrera concludes:
>
>> This example serves as a cautionary tale: even experts can be
>> bewitched by misleading patterns, a danger that has become more acute
>> in the era of “Big Data,” an age in which we now have the ability to
>> gather, process, and analyze massive quantities of data.
>
> When will we learn?
>
> Yours,
> WM
>
> --
> Willard McCarty,
> Professor emeritus, King's College London;
> Editor, Interdisciplinary Science Reviews;  Humanist
> www.mccarty.org.uk


--[2]------------------------------------------------------------------------
        Date: 2021-07-01 11:58:47+00:00
        From: Henry Schaffer 
        Subject: Re: [Humanist] 35.118: phantoms of Big Data

Willard asks, "When will we learn?"
My cynical answer is, "Never."

To elaborate a bit, that will not be until "click bait" goes away - and
it's been with us before clicks - remember "Above the fold?" And our
infatuation with "big data" is just further fueled by the mechanisms of
ML/AI which can handle even bigger data.

I'll end with citing my favorite book/website on correlation
https://www.tylervigen.com/spurious-correlations

--henry


_______________________________________________
Unsubscribe at: http://dhhumanist.org/Restricted
List posts to: humanist@dhhumanist.org
List info and archives at at: http://dhhumanist.org
Listmember interface at: http://dhhumanist.org/Restricted/
Subscribe at: http://dhhumanist.org/membership_form.php