Modelo de clustering para identificar términos que se utilizan conjuntamente para aumentar la cantidad de retweets

The Twitter account of the department of system and computing engineering of the University of Andes is having trouble getting retweets and likes in its tweets. The coordinator of communications says this lack of interaction in Twitter is due to the poor interest from the community towards academic social networks, and not being able to recognize which topics interest the public more than others. This project helps its user understand better which terms, used in a Tweet, generate more retweets based on the topic being displayed in the tweet. This is achieved through prediction, clustering, and a display based on dashboards, that analyzes the behavior of all historical tweets from the department. The solution consists of a pipeline that extracts, cleans, and analyzes tweets to determine which words receive better retweets, based on the topic in which they are used. This solution is deployed in the form of dashboards, each with a different visualization. One with general information from the account, like retweets and likes by date, another which recapitulates retweets by hashtags, likes, and topics, and finally a visualization where you can see which words generate more retweets based on the topic being spoken. This project manages to identify which words have been historically receiving more retweets, based on topics over the years. It also displays information about the state of the tweets from the account throughout time in a straightforward way.

• Autores: Beltrán Ochoa, Santiago
• Tipo de recurso: Trabajo de grado de pregrado
• Fecha de publicación: 2020
• Institución: Universidad de los Andes
• Repositorio: Séneca: repositorio Uniandes
• Idioma: spa
• OAI Identifier: oai:repositorio.uniandes.edu.co:1992/51484
• Acceso en línea: http://hdl.handle.net/1992/51484
• Palabra clave: Redes sociales en línea; Medios de comunicación de masas; Clusters (Sistemas computacionales); Redes sociales; Ingeniería
• Rights: openAccess
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/