The exponential growth of the travel industry, due to the globalization phenomenon, has recently made this sector the focus of several studies. Especially, travel demand forecasting has become one of the most significant issues for the players within the tourism sector: travel agencies and tour operators need to forecast the future demand and plan their offer for the following year. Moreover, due to the financial crisis, predictions need to be as accurate as possible to prevent extra costs and wastes and new machine learning techniques are required to restrict mistakes at least. Alongside, the exponential increasing of data available in the web has given rise to the phenomenon of information explosion, which has completely disrupted the tourism industry. Despite the difficulty of organizing a trip, the growth of the ICTs and the emergence of Internet has allowed the users to access to travel information in a faster and self-service way, resulting in a completely new way for travellers to plan their trips. Additionally, the expansion of internet platforms like social networks, blogs and forums drives the customers to interact with other tourists to have recommendations and share their experiences, leading the players to use this online information to forecast. Starting from these terms, the objective of this thesis is to understand the predicting power of Google Trends and travel blogs. The project, alongside with a graphical analysis, predicts tourists’ arrivals to Italy from eight international countries with heterogeneous sources of data as travel blogs, Google Trends and Google Correlate, using multiple linear regression models and neural networks. Overall, the conclusion of the research is that online data can be used as a support for tourism demand forecasting together with other variables, but the results have to be investigated from country to country.
La crescita esponenziale del settore turistico, dovuta al fenomeno della globalizzazione, ha reso questo settore di rilevante importanza tanto da essere al centro di numerosi studi. Soprattutto la previsione della domanda di viaggio è diventata una delle questioni più importanti per gli attori in gioco. Le agenzie di viaggio e gli operatori turistici devono prevedere la domanda futura e pianificare l’offerta per l'anno successivo. Inoltre, a causa della crisi finanziaria, le previsioni devono essere il più accurate possibile per prevenire costi maggiori e sprechi. Nuove tecniche di machine learning sono quindi necessarie per limitare al minimo gli errori. Parallelamente, l’esplosione di informazioni dovuta all'aumento esponenziale dei dati disponibili nel web, ha completamente cambiato l'industria del turismo. Nonostante la difficoltà di organizzare un viaggio, la crescita delle ICT e la nascita di Internet hanno permesso ai viaggiatori di accedere alle informazioni in modo veloce e fai-da-te, dando vita a un modo completamente nuovo di pianificare i viaggi. Inoltre, l'espansione di piattaforme internet come social networks, blog e forum spinge il cliente a interagire con altri viaggiatori per avere consigli e condividere la propria esperienza, portando gli operatori turistici a prevedere tramite le informazioni online. Partendo da questi termini, l'obiettivo di questa tesi è quello di comprendere il valore di previsione di Google Trends e dei blog di viaggio. Il progetto, insieme ad un’analisi grafica, prevede gli arrivi dei turisti in Italia da otto paesi internazionali utilizzando modelli di regressione lineare multipla e, dove necessario, reti neurali. Sono state utilizzate fonti di dati eterogenee come blog di viaggio, Google Trends e Google Correlate. Questa ricerca conclude che i dati online possono essere utilizzati come supporto per la previsione della domanda turistica insieme ad altri strumenti, ma i risultati vanno interpretati diversamente per ogni paese.
Tourism demand forecasting through Google Trends and travel blogs
ALBERO, MARILENA
2017/2018
Abstract
The exponential growth of the travel industry, due to the globalization phenomenon, has recently made this sector the focus of several studies. Especially, travel demand forecasting has become one of the most significant issues for the players within the tourism sector: travel agencies and tour operators need to forecast the future demand and plan their offer for the following year. Moreover, due to the financial crisis, predictions need to be as accurate as possible to prevent extra costs and wastes and new machine learning techniques are required to restrict mistakes at least. Alongside, the exponential increasing of data available in the web has given rise to the phenomenon of information explosion, which has completely disrupted the tourism industry. Despite the difficulty of organizing a trip, the growth of the ICTs and the emergence of Internet has allowed the users to access to travel information in a faster and self-service way, resulting in a completely new way for travellers to plan their trips. Additionally, the expansion of internet platforms like social networks, blogs and forums drives the customers to interact with other tourists to have recommendations and share their experiences, leading the players to use this online information to forecast. Starting from these terms, the objective of this thesis is to understand the predicting power of Google Trends and travel blogs. The project, alongside with a graphical analysis, predicts tourists’ arrivals to Italy from eight international countries with heterogeneous sources of data as travel blogs, Google Trends and Google Correlate, using multiple linear regression models and neural networks. Overall, the conclusion of the research is that online data can be used as a support for tourism demand forecasting together with other variables, but the results have to be investigated from country to country.| File | Dimensione | Formato | |
|---|---|---|---|
|
FinalThesis_MarilenaAlbero.pdf
non accessibile
Descrizione: Testo della tesi
Dimensione
3.67 MB
Formato
Adobe PDF
|
3.67 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/145044