Publication: Herramienta visual para monitorización tweets durante eventos excepcionales
Loading...
Identifiers
Publication date
2016-07
Defense date
2016-07-12
Authors
Advisors
Tutors
Journal Title
Journal ISSN
Volume Title
Publisher
Impact
Export
Abstract
Vivimos en una sociedad en las que las redes sociales juegan un papel muy
importante en nuestras vidas. Estas nos permiten mantenernos en contacto con
nuestros seres queridos, nos dan un lugar en el que expresar nuestros sentimientos,
o compartir lo que nos encontramos haciendo. La popularidad de estas redes está
en auge, con una enorme cantidad de contenido nuevo publicándose cada segundo.
El objetivo de este Trabajo de Fin de Grado es desarrollar una herramienta
que permita dar un signifiado a esta información, analizándola y visualizándola,
de modo que pueda ser útil para los servicios de emergencia. Estos, de momento,
no emplean esta información, y usarla podría suponer un avance en su labor.
Tras un estudio inicial sobre cómo las redes sociales han sido empleadas por
otros investigadores en el ámbito de los servicios de emergencia, así como un estudio
de las herramientas de visualización de tweets existentes, se propone la siguiente
herramienta visual para monitorización de tweets durante eventos excepcionales.
Esta herramienta recupera información de la red social Twitter y la almacena
en una base de datos. Los tweets - estados compartidos en esta red - son analizados
sintáctica y semánticamente y categorizados en función de su relevancia.
Los tweets se muestran sobre un mapa en las coordenadas en las que fueron
escritos y agrupados en clusters - grupos con coordenadas próximas. Cada cluster
es representado por el polígono de menor área que recubra todos los marcadores
del cluster. Se calcula el término más frecuente entre los tweets que pertenecen al
cluster. La categoría de este es la que determina el color del polígono. Este término
más frecuente también se muestra sobre el mapa.
En último lugar, se aplica la herramienta a un caso de uso real: los atentados
de Parí s del 13 de noviembre de 2015, y se presentan una serie de conclusiones.
We live in a society where social networks play a very important role in our lives. They allow us to keep in touch with the people we care about, they give us a place where we can express our feelings, or share what we are doing. The popularity of these networks keeps rising, with a high amount of new content being published every second. This project aims to develop a tool that will make use of all this information, analyze it and visualize it in a way that it can become useful for the emergency systems. These, at the moment, are not using this information, and using it could lead to an improvement in their performance. After a prior study on how social networks have been used by other researchers in the field of emergency systems, as well as a study on the existing tweet visualization tools, a visual tool for monitoring tweets during exceptional events is proposed. This tool retrieves data from the social network Twitter and stores it in a database. The tweets - statuses shared on this network - are later syntactically and semantically analyzed and categorized based on their relevance. Tweets are shown on a map on the coordinates where they were shared, and clustered into groups with nearby coordinates. Each cluster is represented by the polygon with the minimum area that covers all the markers in the polygon. The most used term in each cluster is calculated. The category this term belongs in determines the color of the polygon. This most frequent term is also shown on the map. Lastly, the tool is tested using a real use case: the Paris attacks of November 13th, 2015, and some conclusions are presented.
We live in a society where social networks play a very important role in our lives. They allow us to keep in touch with the people we care about, they give us a place where we can express our feelings, or share what we are doing. The popularity of these networks keeps rising, with a high amount of new content being published every second. This project aims to develop a tool that will make use of all this information, analyze it and visualize it in a way that it can become useful for the emergency systems. These, at the moment, are not using this information, and using it could lead to an improvement in their performance. After a prior study on how social networks have been used by other researchers in the field of emergency systems, as well as a study on the existing tweet visualization tools, a visual tool for monitoring tweets during exceptional events is proposed. This tool retrieves data from the social network Twitter and stores it in a database. The tweets - statuses shared on this network - are later syntactically and semantically analyzed and categorized based on their relevance. Tweets are shown on a map on the coordinates where they were shared, and clustered into groups with nearby coordinates. Each cluster is represented by the polygon with the minimum area that covers all the markers in the polygon. The most used term in each cluster is calculated. The category this term belongs in determines the color of the polygon. This most frequent term is also shown on the map. Lastly, the tool is tested using a real use case: the Paris attacks of November 13th, 2015, and some conclusions are presented.
Description
Keywords
Redes sociales, Twitter, Tratamiento de la información, Análisis de datos