http://www.academicproductivity.com/2008/who-needs-theories-when-one-has-lots-of-data/ A survival guide for the 21st century researcher Wed, 01 Feb 2012 09:37:35 +0000 hourly 1 http://wordpress.org/?v=3.3.1 http://www.academicproductivity.com/2008/who-needs-theories-when-one-has-lots-of-data/comment-page-1/#comment-43547 John Hunter Sun, 02 Nov 2008 21:58:36 +0000 http://www.academicproductivity.com/2008/who-needs-theories-when-one-has-lots-of-data/#comment-43547 This is an interesting area to examine. I agree that the new tools allow for new strategies. However, I do not believe this will replace the scientific method. Those that use these new tools well will be able to find interesting correlations which can then lead to new insight - by exploring what is going on that they could not have even noticed before. This is an interesting area to examine. I agree that the new tools allow for new strategies. However, I do not believe this will replace the scientific method. Those that use these new tools well will be able to find interesting correlations which can then lead to new insight – by exploring what is going on that they could not have even noticed before.

]]> http://www.academicproductivity.com/2008/who-needs-theories-when-one-has-lots-of-data/comment-page-1/#comment-43239 London Photographer Wed, 29 Oct 2008 13:04:49 +0000 http://www.academicproductivity.com/2008/who-needs-theories-when-one-has-lots-of-data/#comment-43239 But when do theories actually start driving behaviour and therefore trends in data?!! So for example, is google's theory that the importance of a site should be measured on the number of backlinks based on data or is it in fact driving mass backlinking behaviour that results in data that backs up this theory? But when do theories actually start driving behaviour and therefore trends in data?!! So for example, is google’s theory that the importance of a site should be measured on the number of backlinks based on data or is it in fact driving mass backlinking behaviour that results in data that backs up this theory?

]]> http://www.academicproductivity.com/2008/who-needs-theories-when-one-has-lots-of-data/comment-page-1/#comment-30828 drtaxsacto Wed, 09 Jul 2008 19:09:59 +0000 http://www.academicproductivity.com/2008/who-needs-theories-when-one-has-lots-of-data/#comment-30828 Hayek wrote in A Counter Revolution of Science that the danger of having lots of data is that people begin to believe the numbers. The audacious arguments about the Petabyte economy has two potential flaws. First, there is an assumption that the algorithms involved in the net will produce intelligent understanding of the dynamics in the system. That is a highly optimistic view of meta-systems and a very low view of the ability of people to infer meaning from even minimal data. Second, there is an assumption that the data collected actually is accurate. As an economist who occasionally struggles with data on what is going in sectors of the economy - that is a dubious premise. Hayek wrote in A Counter Revolution of Science that the danger of having lots of data is that people begin to believe the numbers. The audacious arguments about the Petabyte economy has two potential flaws. First, there is an assumption that the algorithms involved in the net will produce intelligent understanding of the dynamics in the system. That is a highly optimistic view of meta-systems and a very low view of the ability of people to infer meaning from even minimal data. Second, there is an assumption that the data collected actually is accurate. As an economist who occasionally struggles with data on what is going in sectors of the economy – that is a dubious premise.

]]> http://www.academicproductivity.com/2008/who-needs-theories-when-one-has-lots-of-data/comment-page-1/#comment-30810 Zdenek Tue, 08 Jul 2008 17:37:01 +0000 http://www.academicproductivity.com/2008/who-needs-theories-when-one-has-lots-of-data/#comment-30810 The US financial meltdown shows the dangers of relying on data without a theory. Data have shown that never in the past real estate values decrease in all important regions of US. Wall Street fed enormous amount of data into empirically driven models to show CDOs (now known as toxic waste) were low risk. They lacked a theory that markets go down when bubble burst ... price on that mistake: one trillion USD. The US financial meltdown shows the dangers of relying on data without a theory. Data have shown that never in the past real estate values decrease in all important regions of US. Wall Street fed enormous amount of data into empirically driven models to show CDOs (now known as toxic waste) were low risk. They lacked a theory that markets go down when bubble burst … price on that mistake: one trillion USD.

]]>