Geolocation Data and Sentiment Analysis Combined with Deep Learning for Tourism Destination Management

Abstract

The length of the data makes sentiment analysis on social media platforms like Twitter challenging, typographical errors, acronyms, and special characters. There are several applications for the underlying issue with social media sentiment analysis. An important problem in the tourism sector is the characterization of fluxes, hence sources of geotagged data have already showed promise for geographic research on the sector. The article describes a technique for determining how the general population feels about Cilento, a well-liked vacation spot in Southern Italy. Our strategy is based on a freshly assembled corpus of travel-related tweets.

We intend to present and evaluate In order to describe the spatial, temporal, and demographic tour- ist flows throughout the enormous area of this rural tourism sector and along its coasts, we used a deep learning social geode framework. To distinguish and evaluate the sentiment, we used two specially trained word-level Deep Neural Networks and two character-level Deep Neural Networks. Contrary to many current datasets, our method does not automatically assess the true attitude implied by texts or hashtags. To improve the correctness of the dataset and demonstrate the efficiency of our system, we manually annotated the entire set.