Web crawler research methodology

In economic and social sciences it is crucial to test theoretical models against reliable and big enough databases. The general research challenge is to build up a well-structured database that suits well to the given research question and that is cost efficient at the same time. In this paper we focus on crawler programs that proved to be an effective tool of data base building in very different problem settings. First we explain how crawler programs work and illustrate a complex research process mapping business relationships using social media information sources. In this case we illustrate how search robots can be used to collect data for mapping complex network relationship to characterize business relationships in a well defined environment. After that extend the case and present a framework of three structurally different research models where crawler programs can be applied successfully: exploration, classification and time series analysis. In the case of exploration we present findings about the Hungarian web agency industry when no previous statistical data was available about their operations. For classification we show how the top visited Hungarian web domains can be divided into predefined categories of e-business models. In the third research we used a crawler to gather the values of concrete pre-defined records containing ticket prices of low cost airlines from one single site. Based on the experiences we highlight some conceptual conclusions and opportunities of crawler based research in e-business. -- e-business research ; web search ; web crawler ; Hungarian web ; social network analyis