Social scientists rely on primary data gathered by experiments and surveys or secondary data
supplied by statistical offices. Recent technological progress has enabled practitioners to utilise
novel, and very large, datasets, often revealing, for the first time, information on individual
choices and their preferences. In this study, our aim was to utilize an Internet Census conducted
by [7] as a case study of alternative data source utilisation for the social sciences. The Internet
Census contains regular IP probes of the entire IPv4 IP space in 2012 (251 · 109 observations).
A similar census was conducted by [5] to conduct research on the stability of the Internet. From
a social science perspective, each variation in on-line or off-line activity is actually a revelation
of an individual’s preference for access to the Internet which can, by data-linking, be associated
with a particular time and geo-spatial location. Whilst an individual measure of the activity
at an IP-Address alone does not provide enough information for useful scientific conclusions,
in the aggregate over time and space, such measurements reflect the cultural habits of Internet
access around the globe, revealing potentially fascinating and important use- and preferencepatterns.
There is a growing number of scientific publications that make use of alternative data sources.
For instance, [9] use satellite images of night time light intensity to analyse regional favouritism
of political leaders, whilst [10] study the occurrence of key words in documents as a proxy for