Seungho Yoo, University College London
Introduction
Finding the relationship between Social Network Service such as Twitter and urban protests has been regarded as the important part for understanding the recent rise of the urban protest. Geotag data, especially attached on Twitter, has been focused to capture the linkage of two different dimensions: online and offline, and considered as the fundamental source to identify the linkage rather than just introducing the locations of Twitter users. Previous researches have only focused on the distribution and the density of the geotag data visualising its timely locations as the points on the map during protest periods, however; it has shown a limitation to figure out the dynamic and complex interaction between online and offline whilst many spurious assumptions and expectations have increased. This paper examines the characteristics of contemporary protests with a set of multiple analysis of Twitter data that includes not only spatial pattern of the geotag data, but also keyword analysis, sentiment analysis and the volume of tweets to identify the hidden intersection between Twitter and the urban unrest. It develops with the study of the European anti-austerity protests on 14th November, 2012 which happened on many European cities, particularly focuses on the protests in London, Madrid and Rome.
Research Questions
This study tries to answer the following questions.
1) Do the trends of Twitter data identify the trends of the protest in cities during the protest period?
2) Do the interactions between Twitter and the urban spaces have different characteristics depending on the city?
3) Does the set of multiple analysis lead different results to investigate the recent protest?
Data Collection and Methods
Through the Twitter API collecting toolkit, 0.2 million geotag tweets had been collected from 9th November to 22nd November in 2012 over three cities: London, Madrid and Rome. The flow of the amount of data is counted based on one hour to determine the fluctuations and the correlation with the protest. In order to compare the spatial patterns of data between regions and cities, 200m edge hexagon grid is used to illustrate the data over the cities by GIS tool. For the analysis of keyword and sentiment on Twitter, Alchemy API is conducted to provide some insights as to compare the trends on Twitter and the situations in protest places.
Data Analysis
In terms of data analysis, we found some preliminary outcomes. Firstly, the volumes of the data is picked on the eve of 14th November rather than on the day of the protest in all cities. It could argue that the information about the event on Twitter is mostly proliferated before the event, and it results in drawing attention to the protest. Secondly, the tweets are mainly located on the city centers, and commercial areas tend to show high density in contrast to residential areas. It is understood that Madrid had stronger protests than London and Rome by newspapers. When we examine the spatial pattern of the data, high numbers of the tweet, which contain information about the protest, were generated in strong protest places such as Gran Via. In the third, the data of Madrid exhibit the fluctuations of sentiment with the protest was going on, though other cities do not show the sentiment changes.
Conclusion
The purpose of the paper is to determine the enhanced possibility of the geotag data with the set of multiple analysis. Through the analysis of the European anti-austerity protests in 2012, the results show that the sentiment and the spatial location of the geotag Twitter data will more correlated when protest is stronger. As the urban protests have been spreading out with the increasing use of Twitter that would stimulate urban protests, the geotag Twitter will steadily be regarded as the essential element of the linkage between online and offline in the digital era.