Web data mining for monitoring business export orientation

Abstract

The World Wide Web (WWW) has become the largest repository of information in the world, providing a data stream that grows at the same time as the scope of the Internet does in society. As with most Information and Communication Technologies (ICTs), its digital nature makes it easy for computer programs to analyze it and discover information. This is why it is being increas­ingly explored as a source of new indicators of technology, economics and development. Web-based indicators can be made available on a real-time basis, unlike delayed official data releases. In this paper, we examine the viability of monitoring firm export orientation from automatically retrieved web variables. Our focus on exports is consistent with the role of internationalization in economic development. To evaluate our approach, we first checked to what extent web variables are capable of predicting firm export orientation. Once these new variables are validated, their automated re­trieval is assessed by comparing the predictive performance of two nowcast models: one considering the manually retrieved web variables, the other considering the automatically retrieved ones. Our results evidence that i) web-based variables are good predictors for firm export orientation, and ii) the process of extracting and analyzing such variables can be entirely automated with no significant loss of performance. This way, it is possible to nowcast not only the export orientation of a firm, but also of an economic sector or of a region. First published online: 16 Mar 2017

Keywords

• automatic indicators
• Big Data
• corporate websites
• export
• monitoring
• nowcasting