Journal: Annals of Tourism Research

Abstract

To improve the quality of life of destination communities in the context of an expanding global economy, tourist destinations must sharpen their competitive advantages through increased productivity. In this study, my team investigate the effect of sectoral productivity on destination competitiveness, using Hong Kong as an example. The competitiveness of a destination is measured by tourism-contributed quality of life relative to that of competing destinations. An autoregressive distributed lag model with error correction mechanism is used to model the relationship between destination competitiveness and various sectoral productivities. The productivity effects of several related sectors are identified through an empirical econometric analysis and the results show that destination competitiveness is more dependent on the productivities of core tourism sectors than those of other sectors.

Key Words: Sectoral Productivity, Destination Competitiveness, ARDL-ECM

Publication Date: Nov 2023

View article in the journal

Collaboration organization: World Travel & Tourism Council

Abstract

Because of the strong impact of the retail tourism, governments and the private sector need to consider the potential of retail tourism in their strategies. This report captures retail tourism’s economic importance, current market landscape, external environment, and challenges. By using a combination of sources, from retail and macro-level economic impact data to the consumer survey and interviews with key industry stakeholders, the report provides an in-depth analysis of retail to uncover the factors that influence retail tourism demand and supply at the global and regional levels.

Publication Date: Sep 2023

View WTTC's official website

Journal: Tourism Economics

Abstract

Recovery forecasts for hotel room demand are critical to managing the crisis of COVID-19. This study employs the autoregressive distributed lag error correction model to generate baseline forecasts of hotel room demand for Hong Kong followed by compound scenario analysis to optimize forecasts considering the pandemic’s impacts. The COVID-19 Travelable Index is designed to group source markets by their pandemic situations, vaccinations, policy responses, and health resilience. To capture pandemic-related uncertainty, this study presents three scenarios describing recovery patterns based on trough duration, the quarter for lifting travel restrictions, and the quarter for returning to baseline forecasts. Hotel demand forecasts geared toward each source market are analyzed, revealing strategies to help hotel businesses manage this crisis.

Key Words: Hotel Room Demand, COVID-19, COVID-19 Travelable Index, ARDL-ECM, Clustering

Publication Date: Aug 2021

View article in the journal

Abstract

Credit card fraud is a major source of bank losses. Accurately detect the fraudulent transaction behaviors is considered as the most important aspect of a successful solution. However, evolution of the fraudulent behaviors is a norm. In response to the different classification cost regarding the false positive and false negative cases, pragmatic evaluation metrics help the bank to meet business objective. This research proposes a comprehensive data-driven approach that covers data pre- processing, feature engineering, unsupervised and supervised learning, evaluation with traditional and newly created metrics to find optimal solution for credit card fraud detection. In the experiment of six machine learning model and an autoencoder deep-learning model, the performance is compared and the impact of rebalance treatment, feature engineering methods are evaluated.

Key Words:Credit Card Fraud Detection, Random Forest, Decision Tree, K-Nearest Neighbor, Logistic Regression, Support Vector Machine, Naïve Bayesian, Autoencoder, Deep Learning, Evaluation Metrics

Completion Date: May 2024

Abstract

To address the issue of accurate prediction of ride-hailing service demand, this research adopts the spatial-temporal graph convolution network (STGCN) approach to forecast future ride-hailing service demand of 16 segmented regions in the city of Hangzhou, China. After the step of preprocessing, the data is input into the STGCN, which consists of two spatio- temporal blocks and an output layer to capture both spatial dependencies and temporal dynamics. After the experiment, it is concluded that the research design and methodology are scientifically sound. The prediction is proven to be accurate.

Key Words:Spatial-temporal Graph Convolution Network, Ride-hailing Demand

Completion Date: Jul 2023

Abstract

This is a task of sentiment detection based on customers' review text data. The sentiment target is denoted as positive, neutral and negative. The feature is a composite of customer's textual comments. A couple of machine learning models are implemented, as baseline models. For comparison, finetuned BERT model a transformer-based model is conducted. Dissimilar text preprocess and tokenization methods are adopted in response to the need of different models. In the explanatory analysis, world could method is applied to visualize the most frequent words in the customers' comment. It is conclude that the finetuned BERT model has achieved the best performance in all metrics surpassing all other machine learning models in this experiment.The experiment is sectioned in the following parts as follows:

A - Introduction

B - Setup the platform and shape the data

C - Preliminary text preprocessing for Word Frequency Explanatory Analysis

D - Split the data into train set and validation set

E - Machine Learning Models

F - Transformer-based deep learning model

Key Words: Sentiment Detection, NLP, BERT, SVM, Decision Tree, XGBoost, Naïve Bayes

Completion Date: Oct 2024

Summary

On Page 1, historical changes of the category expenses during 2019 to 2023 are illustrated using ribbon charts. In the view of decomposition tree on Page 2, each expense category can be drilled down into a hierarchical breakdown. To understand the portion of all the components for selected financial data, the tree map with slicer on Page 3 is an efficient approach. Finally, original tabular data is presented on Page 4 for reference of the accurate figures in details.

For managerial implication, please refer to the findings and recommendations concluded.

Completion Date: Sep 2024

Sales Data

Sales data visualizaton.

Completion Date: Sep 2024

Database and tasks

Virginia runs a restaurant in Hong Kong specialized in Mediterraneon cuisine. Someone built a simple reservation management database for her. She found the database is lack of proper configuration. More importantly, she needs to translate the data into valuable business insights for management decision making. For such, she executes series of MySQL statements to complete the tasks below:
A: Initial database and tables establishment
B: Relational database configuration refine
C: Daily database management
E: Table management decision making
F: Data backup
G: Management recommendations
1. Promotion can be targeted to the most valuable guests identified.
2. More tables should be reserved for walk-in guests due to higher consumption price.
3. Collect and create a table of waitlist guests for walk-in customer for better estimation of the total demand so as to optimize revenue and table management.
4. Enrich the categories of food in each food type to appeal to more walk-in guests given the walk-in guests are more lucrative customers.

Key constraint: key constraint, domain constraint, referential integrity constraint

Aggregate function: COUNT(), AVG()

Key MySQL statements:
CREATE TABLE, INSERT INTO, UPDATE, ALTER TABLE - ADD CONSTRAINT, ALTER TABLE - ADD COLUMN,
RIGHT JOIN ON,
GROUP BY HAVING, ORDER BY DESC,
ON DELETE CASCADE, ON UPDATE CASCADE,
BETWEEN, JOIN, ANY, ALL, CASE, SET, REPLACE, UPDATE

Completion Date: Oct 2024

Tasks

Bike retal regression task

Automobile price regression task

Diabetes Classification Task

Penguins clustering task

Azure Service: Azure Machine Learning Studio, Automated ML, Designer

Completion Date: Nov 2024