Computing Reviews
Today's Issue Hot Topics Search Browse Recommended My Account Log In
Review Help
Search
All data are local : thinking critically in a data-driven society
Loukissas Y., The MIT Press, Cambridge, MA, 2019. 272 pp. Type: Book (978-0-262039-66-6)
Date Reviewed: Nov 25 2019

Two important points of general interest can be drawn from this book, although both points require more effort than most readers are willing to make.

The first point, as the title suggests, is that all data is local. This can be taken to mean that data is (or “are,” which we will get to shortly) always created in a context, and what that data means requires one to understand that context. The idea of data objectively representing facts about the world independent of context is hopelessly naive. This point is important because data downloaded and analyzed is often taken to be factual when there are endless reasons why that is not the case. If one wants to understand the data, and consequently the results, of any analysis, one must understand the context within which the data was created. Further, if two sets of data are combined, they are almost certainly from different contexts; any result is thus, to some degree, suspect. On that point, although I have taken the liberty of rewording, I think the author and I agree. Unfortunately, I come from the perspective of a database designer/data analyst, whereas the author approaches the issue as a data anthropologist or library scientist, which rather severely restricts the conceptual overlap.

The second point, which is a little less intentionally revealed, is that our understanding of data is evolving and has yet to be solidified in a new epistemologically workable understanding. This is suggested beginning with the author’s insistence that “data are local” rather than “data is local.” Granted, “data” is generally plural, with “datum” being the singular form, although you rarely hear “datum is” or “data are” because our understanding of the concept of data has been changing and is continuing to do so. The author offers an alternative: “capta.” “Data” is Latin for “given,” while “capta” is Latin for “taken.” If we are going to play with Latin, I would suggest “veritates,” which is Latin for “facts” and suits my understanding better because I see data as facts (preferably, single-valued facts). Still, it is not at all obvious how these Latin roots advance the conversation. It is a bit like discussing an impending tsunami and fretting over whether the “t” is silent. It misses the point rather badly.

There are many problematic inconsistencies in how we use the word “data.” Are the bits of data in big data larger than the bits of data normally found in data that is not big? It is also very unclear what qualifies as data. Is a map of the world data, or is there data included on the map, such as latitudes, longitudes, and country names? Again, we encounter the conceptual conflict between the designer and the anthropologist. The designer might point out some existing ideas, such as experiment design, mistrust of secondary data, database design, proper entity class construction, and data visualization techniques, which may mitigate some but not all problems. The anthropologist is more likely to point out various problems without offering solutions. To be fair, the author does provide some suggestions at the end; however, they are very vague and not particularly useful.

The author attempts to justify the claim that all data is local through four case studies of local data that may or may not be data at all. The cases involve a local something. But the conclusions may or may not be about data, depending on how you define data. For example, is a map or an image or a video or an audio recording really data? Maybe it is information. Maybe it is knowledge. Maybe it is a representation of data, information, or knowledge with potential representation or fidelity issues. Maybe it is something else entirely. Without clarifying what we mean by data, it is very difficult to add to our understanding of it.

Unfortunately, this book adds more confusion than clarity to the evolving phenomenon of data. If you are an anthropologist of data or a library scientist, you might find it a worthwhile and thought-provoking read. If you are a database designer or data analyst, you will likely find it confusing and frustrating.

More reviews about this item: Amazon

Reviewer:  J. M. Artz Review #: CR146798 (2004-0075)
Bookmark and Share
  Reviewer Selected
Featured Reviewer
 
 
Social And Behavioral Sciences (J.4 )
 
 
Law (J.1 ... )
 
 
Arts And Humanities (J.5 )
 
 
General (A.0 )
 
Would you recommend this review?
yes
no
Other reviews under "Social And Behavioral Sciences": Date
Computers and history
Adman P., Halsted Press, New York, NY, 1987. Type: Book (9789780470208526)
Aug 1 1988
A guide to SPSS/PC+
Frude N., Springer-Verlag New York, Inc., New York, NY, 1987. Type: Book (9789780387913124)
May 1 1988
Relational data base structures and concept formation in the social sciences
Edward E. J. Computers and the Social Sciences 1(1): 29-49, 1985. Type: Article
Dec 1 1985
more...

E-Mail This Printer-Friendly
Send Your Comments
Contact Us
Reproduction in whole or in part without permission is prohibited.   Copyright 1999-2024 ThinkLoud®
Terms of Use
| Privacy Policy