What affects the online ratings of restaurant consumers: a research perspective on text-mining big data analysis

Purpose Despite a significant focus on customer evaluation and sentiment analysis, limited attention has been paid to discrete emotional perspective in terms of the emotionality used text. This paper aims extend general-sentiment dictionary Chinese restaurant-domain-specific dictionary, visualize spatiotemporal trends, identify main emotions that affect customers’ ratings restaurant setting constituents influential emotions. Design/methodology/approach A total 683,610 online reviews downloaded from Dianping.com were analyzed by optimized authors; (joy, love, trust, anger, sadness surprise) explored using multiple linear regression methods. After tracking these review texts, Latent Dirichlet Allocation (LDA) LDA models with term frequency-inverse document frequency as weights find factors constitute Findings The results show it is viable optimize or expand word similarity. findings highlight love anger have highest effect ratings. consumers’ (local characteristics, incorrect food portions unobtrusive location) (comfortable dining atmosphere, obvious local characteristics complete supporting services) are identified. Different previous studies, negativity bias not observed, which poses question whether do culture. Practical implications These can help managers monitor true quality service an area time. Based results, operators better decide aspects they should pay more to; platforms operate manageable webpage settings; consumers easily capture restaurants make purchase decisions. Originality/value study builds upon existing general and, best authors’ knowledge, first provide use for analysis. It also reveals two prominent (love anger) case reviews.