AI and Big Data

Why is this topic important? 
Big data and Artificial Intelligence (AI) offer unprecedented opportunities to monitor, forecast, and manage tourism activities in real-time, overcoming the time lags associated with traditional survey data. These technologies enable the processing of unstructured data (text, images, video) to extract insights about tourist sentiment, behavior, and preferences. As destinations strive to become “smart,” integrating AI into destination management ecosystems is crucial for enhancing visitor experiences, optimizing operations, and ensuring data-driven governance.
 
What has our team done so far? 
Our team has pioneered the use of diverse big data sources for tourism forecasting and analysis. We demonstrated that web traffic data from Destination Marketing Organizations (DMOs) can significantly improve hotel demand forecasting accuracy. We applied spatial-temporal forecasting models (like STARMA) and machine learning techniques (Lasso, Elastic Net) to forecast daily attraction demand using search engine and social media data. In the realm of unstructured data, we utilized deep learning (CNN) on social media posts to monitor tourist experiences and health issues under air pollution. We employed Natural Language Processing (NLP) with Large Language Models (like GPT-4o) to quantify “Government Attention to Tourism” from policy reports. Our team also developed a video analytics framework (ASC) to determine which aesthetic and content features of DMO videos on Facebook best engage viewers. Additionally, we proposed a “Smart Destination Management Model” that integrates AI initiatives into a collaborative ecosystem for urban destinations.

Recommendation

Lee, Eunji; Yang, Yang

2026

JOURNAL OF TRAVEL & TOURISM MARKETING

Yang, Yang; Tan, Karen Pei-Sze; Liu, Yi Vanessa

2026

TOURISM MANAGEMENT

Zhang, Xiaowei; Huang, Xingyu; Yang, Yang; Liu, Wei

2026

TOURISM MANAGEMENT

Related Presentations

COVID19tourism Index and its application in tourism management

University of Perpignan

Perpignan, France

Machine Learning and Artificial Intelligence Research in Tourism and Hospitality

University of Macau

Macau (Online)

Tourist behavior analysis using online user generated data

Kyung Hee University

Seoul, Korea (Online)

Spatial Analytics

Workshop on Informatics, Data Science, and Economics in Hospitality and Tourism Research

University of Houston, Houston, TX

Related Resources

Bike sharing and tourism impact tool

Tool link: https://uflyy.github.io/bike-sharing/

The tool is an interactive, data-driven dashboard designed to visualize the synergistic relationship between urban micro-mobility and tourism. Grounded in empirical econometric research, the tool uses Chicago as a case study to demonstrate how bike-sharing systems impact the demand and visitor experience of nearby tourist attractions. It features interactive spatial mapping and a predictive policy simulator, bridging the gap between academic research and smart city tourism management.

Rating Adjustment Tool

Tool link: https://uflyy.github.io/rating-adjustment/

 

The Rating Adjustment Tool is an advanced analytical web application designed to standardize hotel online reviews by correcting for “scaling heterogeneity”—the phenomenon where different types of reviewers interpret and use rating scales differently. Powered by a Hierarchical Ordered Probit (HOPIT) model grounded in peer-reviewed academic research , the tool mathematically controls for systematic response biases tied to traveler demographics (such as age and gender) and trip characteristics (such as travel type and reviewer expertise). By offering individual review adjustments, hotel-level aggregate score calculations, and batch CSV processing, the tool effectively translates subjective, raw user ratings into objective, comparable latent scores and standardized 1–5 metrics, ensuring fairer and more accurate hotel evaluations.

This rating adjustment tool is built on the theoretical framework and empirical results from the following paper:

Leung, X. Y., & Yang, Y. (2020). Are all five points equal? Scaling heterogeneity in hotel online ratings. International Journal of Hospitality Management, 88, 102539.

PS-SAT: Predictive Scheduling & Satisfaction Analytics Tool

This web-based dashboard is designed to visualize tourism and hospitality employee satisfaction data across U.S. geographic units at both the City and Metropolitan Statistical Area (MSA) levels. The application provides interactive mapping, subgroup breakdowns, and (where applicable) legislative pre– and post–comparisons.

Key Features:

  • Dual-Level Geographic View: Users can switch between City-Level (point markers) and MSA-Level (polygon map) visualization.
  • Interactive Search & Selection: Users can search locations dynamically and select them either via the map or the dropdown search interface.
  • Subgroup Breakdown Analysis: Satisfaction scores are displayed by:
    • Job tenure
    • Business model
    • Skill level
    • Front-of-house vs. back-of-house
  • Legislative Effect Comparison: For locations with valid policy-period data, the dashboard displays Pre-Law vs. Post-Law satisfaction comparisons.
  • Visual Encoding of Sample Size: City marker size scales with sample size (N), allowing immediate identification of data-rich locations.

Link: https://uflyy.github.io/ps-tool/