Analysis and Forecast of Post-Pandemic Tourism Recovery in Peru using Time Series modeling

Villalba T.; Soto V.C.; Gutierrez M.A.; Jauregui V.S. · Proceedings of the 2025 IEEE 5th International Conference on Advanced Learning Technologies on Education and Research, ICALTER 2025 (2025) · DOI: 10.1109/icalter69698.2025.11355019

Resumen

The COVID-19 pandemic caused a significant decline in global tourism activity, affecting cultural destinations such as Cusco, Perú. To analyze this situation, this study developed a comprehensive data science model. The objective was to analyze and forecast the recovery of tourism activity in Cusco using historical data on tourist ticket sales from 2009 to 2024. A data science model was developed that incorporated data preprocessing, exploratory analysis, and univariate prediction using the Prophet and Python libraries. Model performance was evaluated using indicators such as the mean absolute error (MAE), root mean square error (RMSE), and mean absolute scaled error (MASE), which assess the accuracy of time series predictions. The results revealed an uneven postpandemic recovery pattern, with domestic tourism recovering more quickly and consistently, while international tourism showed a slower and more volatile recovery behavior. Overall, the tourism sector exhibited a gradual upward trend but remained below pre-pandemic levels. The forecasting model achieved satisfactory accuracy, with an MAE of 1,408.67, an RMSE of 1,707.88, and an MASE of 0.5. The visits range from 10,000 to 50,000, and the observed errors indicate a relatively low deviation, which suggests reliable performance. These results demonstrate the effectiveness of forecasting in modeling tourism recovery and its potential contributions to strategic planning and government policy development. It also showed the uneven recovery of the sector and the urgent need to be prepared in periods of uncertainty.

Instituciones: Universidad Nacional de San Antonio Abad del Cusco; Universidad Andina del Cusco