O principal objetivo deste estudo é entender como as plataformas online influenciam a escolha do cliente no Canal dos Restaurantes. Foi realizada uma análise em Portugal que inclui 487 comentários de diferentes usuários. O modelo de máquina de vetor de suporte proporcionou resultados interessantes devido à modelagem efetiva da pontuação do usuário, com base nos recursos coletados das plataformas, com um erro médio de percentagem absoluta de 10,15%. Várias variáveis foram analisadas, nomeadamente a plataforma em análise, a localização, o número de fotos e o dia da semana. A característica mais relevante observada foi o ranking que cada consumidor atribuiu a cada restaurante. No estudo realizado nesta dissertação verificamos que o Ranking pesa 24% na escolha do consumidor no seu restaurante. De seguida, verificamos que a Plataforma escolhida e as fotografias que são mostradas nesta contribuem 19% e 18% respetivamente. Através da análise feita verificámos que a plataforma Zoomato é a mais usada pela nossa amostra. Desta forma, verificamos que a resposta à nossa pergunta inicial realização desta dissertação: "Como é o comportamento do consumidor e como é influenciado por estas plataformas?" consegue ser respondida mais à frente através da análise dos resultados.

The main objective of this study is to understand how online platforms influence customer choice in the field of restaurants. To this end, various social networking platforms were analyzed, namely TripAdvisor, Zoomato, Facebook and Yelp, these being the most commonly used by consumers when looking for information that will help them to choose a restaurant. The platforms studied are among the most widely used by consumers when searching for information that may guide their next purchase. This analysis was conducted in Portugal, and includes 487 reviews from different users. The support vector machine model provided interesting results due to the effective modelling of the user score, based on the features collected from the platforms, with a mean absolute percentage error of 10.15%. Several variables were analyzed, namely the platform, the location, the number of photos and the day of the week. The most relevant characteristic observed was the ranking that each consumer attributed to each restaurant. Through the results that will be presented we see that the Avg.Rate shows a 24% contribution to other users he most relevant characteristic observed was the ranking that each consumer attributed to each restaurant. The variables "Platform" and "nr.photos" contribute 19% and 18% respectively and therefore do not represent a big difference. It can be confirm that the values coincide with Zoomato, which was the Platform most used by our sample, having studied the following variables: Platform, Location, District, Website, Rank, Cuisine, Avg. Rate, Number of photos, Day and Month. In this way, it was verified that the answer to our initial motivating question, "What is consumer behaviour like and how is it influenced by these platforms?" confirms that consumers won’t go to a new restaurant before checking the ranking that has been attributed to it on the various analyzed platforms.