A Latin American Demographic Map
of Internet Activities
The World Wide Web is not one specific service, but a medium through which different services can be offered. Websites can be built to serve a wide variety of purposes. Some of these websites (such as newspapers and radio) are derived from traditional media models and is facilitated by the re-purposing of traditional media content. Other websites (such as web auctions) do not have an equivalent counterpart in traditional media.
By virtue of the multiplicity of services, the World Wide Web is able to attract many people who come for different purposes. In the history of traditional media, a new medium starts with a few large players who need to appeal to a mass audience in order to growth. As the medium matures, niche players emerge to serve specific segments. In the case of the World Wide Web, the development was compressed into a much faster timeframe. Today, in less than a decade, the World Wide Web has millions of websites that cover every conceivable subject.
In this note, we will present some data about the demographic characteristics of users of different types of websites. We will cite some data from the TGI.net Latina study. The data come from 6,563 Internet users (within the past 90 days) between the ages of 12 and 64 years old in Argentina, Brazil, Chile, Colombia and Mexico. In this study, the Internet users were shown 25 types of websites, and they indicated the ones that they visited most often.
Our interest is not so much in which types of websites are most popular. We know that some types of websites are more popular than others. We were interested the degree to which website types are liked by different segments of the population. Specifically, we were interested in the age/sex segmentation. Our data are represented graphically by a correspondence analysis.
The correspondence analysis yielded the following graphical representation of the age/sex groups: along the horizontal axis, the groups are sorted by age, with the youngest people on the left and the oldest people on the right; along the vertical axis, the females are on the top half and the males are on the lower half. This graphical representation was created purely on the basis of the website preferences of the survey respondents, as the algorithm has no semantic understanding of the meanings of the labels (such as males 12-19). Thus, the natural age/sex structure was revealed by the algorithm purely from the data structure. In the same map, the websites fall into the quadrants where one would expect them to be. This graph represents the data much better than hundreds of words ever could. Although we would like to refrain from editorializing, we cannot help but point out that the reader should look at the position of 'adult entertainment.' Adult ... ?
(Source: TGI.net Latina)
The most accessible description of correspondence analysis is: Michael Greenacre (1993). Correspondence Analysis in Practice. Academic Press: New York.
(posted by Roland Soong, 1/26/2001)
(Return to Zona Latina's Home Page)