Proyecto final de Grado: Twitren
Archivos
Fecha
2012
Título de la revista
ISSN de la revista
Título del volumen
Editor
Resumen
"La presente investigación analiza la factibilidad de desarrollar un sistema que permita predecir el tiempo de arribo de trenes de pasajeros, en un ramal de la provincia de Buenos Aires, Argentina, utilizando información aportada por los pasajeros del tren. Esta información es la geo localización y timestamp de mensajes publicados en la red social Twitter, y emitidos desde las formaciones en movimiento.
Se desarrolló un modelo de simulación que contempla el tren en movimiento, y la emisión de Tweets desde el tren y los alrededores de la vía. Mediante la utilización de puestos de control, se junta la información de esas señales, para predecir la ubicación del tren y por ende su horario estimado de llegada a las distintas estaciones siguientes.
Se analizaron las variables claves a tener en cuenta para el éxito del modelo, entre las cuales se incluyen costos, difusión, tecnologías, etc.
Las distintas simulaciones realizadas permitieron calibrar el sistema, y determinar los valores críticos de las variables involucradas, que dimensionan el modelo y lo contrastan con los valores encontrados en la realidad".
"This investigation studies the feasibility of developing a system capable of predicting the time of arrival for passenger trains in Buenos Aires, Argentina, by using information provided by those passengers (location and timestamp of short messages posted by users in the social network Twitter, from within moving trains). A simulation model was developed, which included the moving train, the emission of messages in Twitter by passengers, and from nearby people. Making use of checkpoints, the information from Twitter was processed to estimate the train's current location, and thus tell its estimated time of arrival to the subsequent stations in the system. Different related-variables were studied, such as costs, awareness generation, technology, etc. Different simulations carried out allowed us to tune the system, determine the critical values of involved variables, and size the model, while becoming aware of the existing gaps between these variables and real-world ones."
"This investigation studies the feasibility of developing a system capable of predicting the time of arrival for passenger trains in Buenos Aires, Argentina, by using information provided by those passengers (location and timestamp of short messages posted by users in the social network Twitter, from within moving trains). A simulation model was developed, which included the moving train, the emission of messages in Twitter by passengers, and from nearby people. Making use of checkpoints, the information from Twitter was processed to estimate the train's current location, and thus tell its estimated time of arrival to the subsequent stations in the system. Different related-variables were studied, such as costs, awareness generation, technology, etc. Different simulations carried out allowed us to tune the system, determine the critical values of involved variables, and size the model, while becoming aware of the existing gaps between these variables and real-world ones."
Descripción
Palabras clave
SIMULACION, FERROCARRILES, REDES SOCIALES