How do consumer perceptions translate into measurable behavior in service industries? We use three complementary methods to find out.
Online reviews, social media posts, booking platform metadata — consumers leave a trail of genuine intent that surveys rarely capture.
Our group scrapes, cleans, and structures hundreds of thousands of reviews per project. We apply sentiment analysis, semantic network analysis, and topic modeling to surface the questions hospitality and tourism customers actually ask — not the questions researchers think to put on questionnaires.
VADER, lexicon-based, and transformer-based classification of multilingual reviews
Co-occurrence matrices, centrality, community detection in R and SPSS
Citation network analysis, research trend mapping with VOSviewer
R + tm/textstem pipelines for hospitality review datasets at scale
Self-reported preference is unreliable. Gaze is honest. We measure where users actually look — on menus, ads, web interfaces — and compare that to what they say they noticed.
BAR LAB is building wtbd.site — a research-grade web-based eye-tracking platform using WebGazer.js (TensorFlow Facemesh) and Cloudflare's edge infrastructure. Researchers upload stimuli, define areas of interest (AOIs), recruit participants, and download AOI dwell-time and saccade data — all without specialized hardware.
Webcam-based gaze tracking with TF Facemesh, calibration grid, and AOI export
Dwell time, fixation count, scan path entropy across regions of interest
Menu layouts, ad placement, web wireframes — controlled visual stimuli
Per-participant + aggregate visualizations exported to PNG/CSV
Strong theory survives empirical contact. We extend, modify, and validate behavior frameworks against real-world data from Asian and emerging markets — places where the original models often don't quite hold.
Theory of Planned Behavior, Technology Acceptance Model, UTAUT2, Post-Acceptance Model — these are our starting points. We introduce constructs (cultural authenticity, halal-friendliness, K-Culture diffusion, financial well-being) and test the modified models with structural equation modeling on cross-cultural samples.
Adding moral norms, cultural authenticity, perceived value to standard TPB
Technology adoption in emerging-market hospitality services
AMOS, SmartPLS, factor analysis for measurement model validation
Quantitative validation paired with qualitative pilot/post-hoc work
Text mining (tm, textstem), bibliometrix, openxlsx, ggplot. The lab's primary statistical workhorse.
Factor analysis, SEM. Reliable for cross-cultural samples and reviewer expectations in hospitality journals.
Variance-based SEM for theoretical models with smaller samples or formative constructs.
Bibliometric network visualization — co-citation, co-authorship, keyword co-occurrence.
Edge-deployed SQLite for the eye-tracking platform and BAR LAB's own infrastructure.
Facemesh-based gaze tracking in the browser via WebGazer.js. No specialized hardware required.