Please use this identifier to cite or link to this item:

http://hdl.handle.net/10609/81516
Title: Análisis predictivo de datos abiertos sobre el uso turístico del servicio de alquiler compartido de bicicletas de Nueva York. Una perspectiva desde la Ciencia de Datos
Author: Jiménez Gómez, Carlos Eugenio
Director: Parada Medina, Raúl
Tutor: Casas Roma, Jordi
Others: Universitat Oberta de Catalunya
Keywords: Smart Government
gobierno inteligente
machine learning
data science
Issue Date: Jun-2018
Publisher: Universitat Oberta de Catalunya
Abstract: In this work we analyze an open data set from the Data Science perspective. The data set is on bicycle trips of non-subscriber users (usually tourists and visitors) of the New York City bike-sharing service. The local government organizational context we will be taken into account for final proposals. Two are the main goals: a) on one hand, we are looking for answers related to a segmented use of the service, and b) on the other hand, from a machine learning perspective, we want to find the best model that predicts the bicycle daily demand, taking into account weather (precipitations) and events (non-working days). We train and test seven different algorithms models, that later we use to predict the demand. The model that best results has obtained in predictions is the Artificial Neural Network. Conclusions highlight, the fact that organizational goals have to guide the full Data Science process. Results support the fact that data pre-processing and processing stages are not trivial nor can be generalized, and it can even become strategic. It is also highlighted the fact that literature should include more accurate explanations on the data pre-processing and processing stages, in the same way that does with algorithms. Based on the data analysis, visualizations and machine learning results, some useful proposals are included from both, operational and strategic perspectives. A fully data-oriented public organization has to be a strategic goal if consolidated advances towards Smart Government are planned.
Language: Spanish
URI: http://hdl.handle.net/10609/81516
Appears in Collections:Bachelor thesis, research projects, etc.

Share:
Export:
Files in This Item:
File Description SizeFormat 
PPT_Jimenez_Predict_CitiBikeNYC_TFM.pptxDocumento PowerPoint con los aspectos destacados del Trabajo Final de Máster2.6 MBMicrosoft Powerpoint XMLView/Open
Video_Presentac_Jimenez_Predict_CitiBikeNYC_TFM.movVideo de presentación con los aspectos destacados del Trabajo Final de Máster448.93 MBVideo QuicktimeView/Open
carlosjgTFM0618memoria.pdfMemoria del TFM7.8 MBAdobe PDFView/Open
carlosjgTFM0618presentación.pdfPresentación PDF del TFM1.35 MBAdobe PDFView/Open

This item is licensed under a Creative Commons License Creative Commons