Economic growth is a crucial indicator for any country's economy. Tourism plays a vital role in the national economy of Indonesia with its multiplier effect, influencing the development of other sectors. Aspect-based sentiment analysis has emerged as a popular approach for understanding the opinions of tourist reviews. However, most aspect-identification algorithms are based on sentence dependencies, which are computationally expensive for longer texts. Therefore, in this paper, we studied Local Sentiment Aggregation (LSA) as an efficient method to extract aspects and classify sentiment polarities of these aspects. LSA introduces aggregation window-based sentiment learning (AW), a mechanism based on embeddings for neighbouring words. The results of our research demonstrate that the LSA model performs well on tourism reviews, as shown by the model's accuracy, which reached 92.48% and an F1 score of 87.40%. copy; 2023 IEEE.