Generation of Domestic Tourism Travel Time Series Using Big Data From Mobile Phone Data

Additionally, traditional methods of measurement are resource-heavy and time-consuming. This paper proposes rules and processes for deriving time series of domestic tourism flows for overnight trips from passive mobile phone data (PMD) in Chile. The data generated are monthly, covering the period from January 2015 to March 2019. To verify the data's accuracy, we examined the tourism time series of Valparaíso, the most visited domestic region in the country. This analysis showed significant correlations with other time series. To gather insights about the time series, forecasting models are built using four different techniques. The SARIMA model provides a forecast with MAPE ≤ 6.0% and highlights the importance of is lag 1 and 12 for several models. In conclusion, the constructed time series displays notable correlations and matches with data from other sources, while the forecasting models behave as expected for tourism data. This indicates that the method is promising for generating reliable domestic tourism time series, saving costs and time, and increasing the frequency of updates.