Please use this identifier to cite or link to this item: https://hdl.handle.net/2445/177707
Big Data Marketing: Transformación de datos y análisis de series temporales
Journal Title
Authors
Director/Tutor
Journal ISSN
Volume Title
Related resource
Abstract
[en] Currently, data analysis is entirely incorporated in marketing, and Big Data processes are increasingly standard in these sectors. The main purpose of this work is to apply Big Data techniques and time series analysis to obtain tangible results for a marketing project. The methodology used follows three steps: first, massive data mining from Google as well as internal data; secondly, the transformation of this data and the creation of multiple data bases; and thirdly, an exhaustive analysis of all the data obtained. In terms of results, search patterns in Google have been identified and classified in accordance with the searcher’s intentions. This, together with data mining, has resulted in the collection of multiple variables which are crucial pointers for future marketing campaigns. An example of this would be the percentage of people who enroll after having requested information (lead), or which Google searches lead to higher enrolment numbers. Time series have also been generated from the variables and the correlation among them has been studied. Very interesting correlations have been found such as a 0,88 in the count of users requesting information and the count of users who click-through to a website after having carried out a transactional search (transactional click-throughs). Finally, an analysis of the time transactional click-through series and a 30 week prediction based on auto-regression models for mobile media, have been carried out.
Description
Treballs Finals de Grau de Matemàtiques, Facultat de Matemàtiques, Universitat de Barcelona, Any: 2020, Director: Josep Vives i Santa Eulàlia
Subject (English)
Citation
Collections
Citation
LLENAS PUIGDEMONT, Jesús. Big Data Marketing: Transformación de datos y análisis de series temporales. [consulted: 7 of August of 2026]. Available at: https://hdl.handle.net/2445/177707