Hong Kong (Hong Kong Special Administrative Region) – November 6, 2019 (travelindex.com) – Thanks to a groundbreaking study by Professor Rob Law of the School of Hotel and Tourism Management (SHTM) at Hong Kong Polytechnic University and his collaborators, restaurateurs now have the opportunity to gain comprehensive insights into the eating habits and behavior of tourists by analyzing online reviews. The researchers developed a data mining framework to analyze over 40,000 restaurant reviews in Sydney and Melbourne, Australia. They uncovered a wide range of tourist preferences and suggest that their method could be extended to other areas of tourism.

Food is a basic need and an important part of our lives. This is especially true on vacation, researchers argue, as food plays a central role in the tourist experience and accounts for about a third of their total spending. It is therefore hardly surprising that tourism professionals demand a deep understanding of tourists' eating habits and preferences in order to optimize their offerings and marketing campaigns. They are interested not only in preferred food types but also in the specific dishes consumed. However, such information is scarce, primarily because traditional research methods like surveys cannot capture this level of detail on a large scale.

Fortunately, popular websites like TripAdvisor and Airbnb encourage their users to share their travel experiences and reviews of restaurants, hotels, and other services. This creates a valuable source of information about tourists' preferences and behavior. The researchers therefore developed a general research model that can process and analyze restaurant reviews and provide insights into the "preferences and subjective opinions of tourists."

To best test their model, the researchers realized they should focus on restaurant reviews in Melbourne and Sydney, Australia's two largest cities. They found that "culinary tourism has become an essential component" of Australia's international marketing strategies and that the country offers a "global cuisine" that combines the foods, flavors, and cooking styles of its large migrant population.

The researchers developed web crawler software to automatically extract reviews. They then applied various preprocessing techniques to transform the content into usable data. For example, a geocoding tool was used to determine the reviewers' locations. Text mining techniques were used to identify relevant words in the reviews, such as different types of food or dishes, for use in subsequent analyses. A previously trained model was applied to the reviews to assess the sentiment expressed and then categorize each sentence as positive, negative, or neutral.

The data collection software was then applied to 40,948 reviews of 2,265 restaurants in Melbourne and Sydney. Location information revealed that reviewers came from 167 of the world's 195 countries. Given this wide distribution, the researchers focused their case study on the 14 countries with the most reviews, ranging from the UK with nearly 11,000 to Germany with fewer than 500. These countries were then categorized as either "Western countries" or "Asian countries."

Italian cuisine was the most popular, accounting for almost 20% of all reviews, followed by Chinese and Japanese. Certain tourist groups, particularly from Italy, India, and France, preferred their respective national cuisines, while Malaysian tourists favored Chinese restaurants and Indonesians Japanese and Thai dishes. Researchers recommend that tourism providers pay special attention to these groups and highlight the availability of their preferred dishes to enhance their dining experience and overall satisfaction.

A preference for one's own country's cuisine did not necessarily translate into satisfaction with the food. While many Italians left positive reviews, numerous guests also expressed negative comments, particularly regarding sauces and pasta. Indian tourists seemed more satisfied with restaurants offering Indian cuisine, especially with dishes such as wine, fish, curry, and lamb. Restaurant managers should consider the opinions expressed in the reviews. Knowing which dishes are preferred and receive the most positive feedback can help them with menu planning and guide different customer groups toward their favorite dishes.

The study also examined which food and drinks were offered in different types of restaurants. Given Australia's reputation for wine production and the growing popularity of wine tourism, the researchers were not surprised to find that wine was consumed in 11 out of 13 restaurant types, particularly those serving Western cuisine. In pubs, beer was almost four times more popular than wine, while cafes were primarily known for light fare such as coffee, desserts, eggs, cakes, and chocolate.

Fewer tourists left reviews for breakfast spots than for other meals. However, when they did leave reviews, they most often ate breakfast in cafes, possibly because their choices are limited by the opening hours of other restaurants. Since cafes are least popular in the evenings, they could focus on offering more lighter options on the breakfast menu to attract more customers at that time of day.

Indian restaurants were least popular at lunchtime, while cafes were most popular in the evening. Restaurant managers should consider their guests' meal preferences and adjust their offerings accordingly. For example, wine, steak, and lamb seem to be very popular in the evening, while salad, burgers, fish, fries, and beer are more suitable for lunch.

Western tourists visited pubs more frequently than Asian tourists, presumably due to their British heritage. As the researchers note, "many countries in Europe and North America are influenced by British culture." There was little difference in popularity of cafés, although Indian tourists visited them least often. Given Western tourists' preference for pubs, tourism professionals could offer "pub experiences" to this target group.

For many tourists, a restaurant's features are just as important as its cuisine when choosing a restaurant. Service was the most important factor, followed by staff and price, while value for money and atmosphere received little mention. Since Western tourists apparently place more value on ambiance than Asian tourists, the researchers recommend that restaurants more popular with Western guests, such as Italian restaurants and pubs, should focus on service quality and other features. Operators of Chinese, Thai, and Vietnamese restaurants, on the other hand, should invest their limited resources in improving food quality, as Asian tourists frequent these types of restaurants more frequently and place less emphasis on ambiance.

Given the comprehensive results, the researchers comment that their data mining framework could be successfully extended to other areas that require a better understanding of tourists' preferences. For example, they suggest using it, after further refinement, to analyze photos of food and restaurants in reviews in order to "gain insights into tourists' experiences and specific interests." Indeed, the potential of such techniques opens up exciting possibilities for understanding consumer attitudes and behavior in all their facets.

Editor's Note: Top25Restaurants.com has developed the world's first AI-powered restaurant rating index and further enhanced the credibility of machine learning by incorporating expert knowledge from the culinary industry. Top25Restaurants.com presents the best restaurants in Asia… click here for more…