Sorry, I thought I attached a pdf. I don't see how to attach to this list so
this is a past of the contents.
Useful tools from other disciplines.
A video clip recently struck me. The clip visualized the dispersal of cell
phones over a month starting near a meat processing plant. The clip
provided strong insight into social contacts and Covid-19 contagion.
The clip is aimed at the business community, but this kind of tool could
provide useful insight for the policy community and for the social and
behavioral science community. It can be found at
https://www.youtube.com/watch?v=F0qAT65V1P0The producer is
http://www.tectonix.com/Techtonix and SPSS are intended only as examples and not as advertisements.
The clip visualized the potential spread of the Covid 19 virus. But the clip
also struck me in other ways. It can speak to social and behavioral
scientists from many fields. This clip shows a tool that has potential to be
adopted and adapted by other disciplines. This clip used unobtrusive
measures and visualization. This approach gives insight not only into
spread of a contaminant but also into spread of fads, ideas, etc.
Different disciplines develop, adapt, and adopt tools to aid perception and
cognition about phenomena. This clip is a specific example of people in
geographically spread places being connected for our food supply.
At the Classification Society, how different disciplines could learn from
each other often struck me. I also saw how very much the same kinds of
analysis tools would develop with the same aims but different names. Some
kinds of exploratory analysis came to be called data mining. Data mining
evolved into big data. Big data evolved into data science. Cluster
analysis, pattern detection, and unsupervised learning were very much the
same tool. Discriminant function, pattern recognition, and supervised
learning were very much the same tool.
Likewise factor analysis and feature selection served similar functions.
Since I retired in 2001, I saw more examples of tools evolving to help other
disciplines. At a joint meeting of the Classification Society and Digital
Humanities, I saw how similar the approaches were when considering text data
such as manuscripts from Scandinavian epics. A little later, I saw how much
useful those text analysis techniques could be in the intelligence
community.
One thing that helped all these tools was the growth in computer power. It
used to take may years of monk labor to generate a concordance of the Bible.
Today once we enter data, it takes mere minutes to generate a concordance.
When Mosteller and Wallace first did a statistical analysis of the
Federalist papers they used frequencies of 18 words. About 2009, I used
Python and SPSS to look at frequencies of 1000 words and it took a couple of
minutes. Someday, I hope to do the same with dyads and triads of words
within sentences.
Around 2000, basically only a few people at MIT were able to handle both
spatial and temporal repeated measures (autocorrelation) simultaneously.
SPSS now has the ability to model both time and spatial data simultaneously.
-----
Art Kendall
Social Research Consultants
--
Sent from:
http://spssx-discussion.1045642.n5.nabble.com/=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
Art Kendall
Social Research Consultants
Locked
