Geo-location of Business – Determinants and Methods of Prediction

Agglomeration as a subject of analysis remains popular – all the time new methods and approaches appear, which is related of course to the availability of data. This phenomenon seems to be totally explained, however, due to its nature there are still some unsolved questions left. Only recently, it became possible to measure knowledge spillovers between firms (Sweeney & Gómez‐Antonio, 2021), so we may expect further studies on the topic. This thesis presents a new look at agglomeration – from the perspective of point pattern analysis. First of all, I analyse agglomeration and related definitions in terms of what they show, mean and measure, and this analysis helps me to find the definition which describes an agglomeration (Marshallian and Jacobian) the best from the point pattern analysis point of view (Chapter 1). To study Marshallian agglomeration, I take the ideas of Sweeney and Gómez-Antonio, who were the pioneers of using the Geyer saturation process for location and agglomeration studies. In the thesis, I am trying to extend their approach by using a hybrid Geyer model and connecting it with a distance-decay issue, which was widely used in agglomeration studies (Chapter 2). With geolocated data of enterprises in Lubelskie voivodeship, I confirm their conclusion that the Geyer saturation process is a good tool for both explaining and modelling location and agglomeration, and with simulated data – I show, that hybrid models perform better than a simple Geyer model and there is an argument for their use in the future. Distance decay issue is also present – it was shown that values of interaction parameters which represent the 'strength' of agglomeration forces decrease rapidly with a distance. The last chapter (Chapter 3) focuses on Jacobian agglomeration and shows currently available methods for analysing it. This chapter is mostly experimental, however, it shows what has to be done and be implemented in point pattern analysis methods in order to be able to reveal the true nature of unobservable connections between different industries' firms.