This week, in a discussion about media ethics, I was introduced to an article by Siva Vaidhyanathan at Slate (2021) addressing concerns about concentration in media ownership after the Telecommunications Act of 1996. The article presents a form of an update, after an earlier work by The Nation, The National Entertainment State, 2006.
This same week, there was an article published at Vanity Fair, “Feels Like a Real Slap in the Face”: There’s New Money in Local News, Just Not at the Papers, in which the August acquisition of Axios by the Cox media brand was contrasted to the effects of layoffs at newspapers under the Gannett name. After the article by Siva Vaidhyanathan, it might serve to inspire some questions about the origins and effects of media ownership in general.
Following some preliminary research with online databases for corporate information, and inspired by the Public Accountability Initiative at LIttleSis, I’ve begun to take a look at using resources at DBpedia for running queries about recent media ownership information in the public sphere. My hope is to be able to display something like a radial graph illustrating the most up-to-date media ownership information available, within any single media field.
Theoretically, such a media ownership graph could be based on a single-instance database of media ownership information, however derived from individual sources – with source references included, for purpose of data quality. Within the data infrastructure for the graph, it could make use of any number of Semantic Web technologies including linked data (e.g RDF) and ontologies for linked data, as vis a vis OWL.
In a first thought, the set of media categories for the graph might include e.g traditional news (newspapers), popular periodicals, Web news sources (e.g Axios), Cable or satellite television networks, broadcast TV or DTV networks, broadcast radio networks, and maybe publishers (books). A category for streaming media networks should also be presented (e.g Showtime/Paramount+)
Initially, my main interest is towards highlighting the ownership relations among Content Companies (to borrow the term from Vaidhyanathan) in forms of mass media (e.g Web, TV, radio, and the magazine shelves at the contemporary store). In the interest of graphing for influence relations and after the work at LittleSis, I’d also like to be able to draw in any corporate board information such that may be available at DBpedia, and to add some more breadth to the containing graph – e.g marketing/advertising companies and media delivery services, as well as information about advertising placement by political campaigns.
Ideally, this could be produced as to illustrate a sense of duration for the data, extensional to any snapshot at an instance of time. Albeit, this feature of the graph would be no less limited to the information available via public sources.
At some point, maybe there will be a nice graphic to share here, in illustration, something like the data visualization example at the D3 web site. Presently, I’m still trying to decide on an architecture for any initial prototype of a visual graphing tool, to such effect – there may not be any really easy alternatives for a cross-platform tool for any number of desktop and/or mobile computing platforms, such that would use SVG graphics and HTML in some visual aspects of presentation. Perhaps the desktop-based tool could then be extended for some single web-based presentation. For this blog article, I mostly wanted to sketch out the idea, at this present point in the design.
As a sort of an illustration, here’s a screen capture from a query with DBpedia, showing four cable service providers in the US. The first company in the list might be in something of a category unto itself, in scale of annual revenue and in number of corporate interlinks, broadly. In a general sense, perhaps this may serve to provide an initial illustration of a methodology for using SPARQL in analysis of media ownership information at DBpedia.
Individual pages at DBpedia would provide further public information, in a generally machine-readable format:
In retrospect, it seems that this single data query may serve to highlight a couple of cable line providers, and Cox. Perhaps it’s not the most socially compelling illustration possible. The query could certainly use some refinement, to sort out the duplicate results in the first column of output. The query itself required some research with DBpedia, to discover any number of linked data properties that might be of use here, given the data set at DBpedia. Perhaps it may serve to illustrate some of the semantics of this approach to a data query.
Some similar searches:
- Television producers (Television)
- Television producers (Television show)
- Television producers (other variations, “Car” and “Chemical industry” excluded for purpose of data quality)
- Media Companies (US) via Wikipedia
- Television Broadcasters (US) via Wikipedia
An introduction to SPARQL is available in the Stardog documentation, with further SPARQL examples available in a 2009 talk by Lee Feigenbaum, at the W3C. (There might be a books list to share, in some later post to this blog)
After discovering that there may be very little information available from federal or state government registrars, for any media corporation formed as a privately traded corporation – contrasted to the public information available via the SEC, for publicly traded corporations – and after taking a look at resources available via DBpedia, I’m confident that there is public data available for a corporate media ownership database, independent the form of any single corporation.
I’ll try to post any updates to the Oort blog at Thinkum.Space, as this project develops over time.