Author(s):
Romero, Elisabet Adeva
Date: 2014
Persistent ID: http://hdl.handle.net/10362/11548
Origin: Repositório Institucional da UNL
Subject(s): Protest; GDELT Global Database of Events, Language, and Tone; Spatial Autoregressive model; Spatial Point Pattern Analysis; Spatial distribution pattern; Spatstat
Description
Dissertation submitted in partial fulfillment of the requirements for the Degree of Master of Science in Geospatial Technologies
Protest in Europe where analyzed to foster an understanding of the distribution and the behaviour of those during from 2000 to 2010 time frame. The main object of this study is to discover if there is a relation between economic, social and other variables available in Eurostat in order to discover a pattern in the protests in Europe. For this purpose, least squared method and spatial point pattern analysis method were applied in the R Software environment. The final output indicates that variables can’t explain a cause-effect relation of protests due to tis behaviour is complex and Europe is an inhomogeneous area. In the other hand, we saw that protest tend to increase mostly when other protest have happened in the past. Protest location are scattered within the European megalopolis, and reveals attraction to some capitals some hot spots patterns are observed. They are mostly located in urban areas, close to the borders with other European countries. The resulting models discovered that protest/events distributions do not imitate an inhomogeneous Poisson process and thus we tried to model the behaviour describing special interaction between locations of protests. The best interaction model was chosen by computing different distances. We analyzed the whole Europe area and due a strong influence of United Kingdom we computed the same model to Germany, France, United Kingdom and Spain. Finally, a step further spatial-temporal analysis was taken only for Spain. This analysis is one of the first analyses set by the recently launched Global Database of Events, Language, and Tone (GDELT), a big free online data base of over 250m events and 300 categories including riots and protests codified from world news sources. After this analysis we recommend, further analysis should contain models that apply border contagion including time.