*3.5. Take-Out Service*

In China, the number of online takeout users accounts for more than 44% of sales, and the scale has exceeded 398 million people (https://www.qianzhan.com/analyst/detail/22 0/200512-65621d53.html (accessed on 2 September 2021)). The take-out markets with its large number of users and rapid growth has generated a huge amount of takeout data. The takeout big data service platform not only helps the governmen<sup>t</sup> supervise the industry, but also creates huge economic and social value. First, it predicts and informs customers of the delivery time, thereby avoiding disrupting consumers' daily plans and helping restaurants establish a good reputation. Second, it helps the take-out enterprise understand consumer demand. Third, the take-out big data platform promotes the transparency of the supply chain, which is conducive to establishing and improving customers' trust. Fourth, the overall running of the city can be clearly understood by analyzing the take-out dataset [66].

Since take-out data involves sensitive private information (the customer's location, preference, bank, identity, and communication), ensuring data security in the take-oout big data platform is a serious challenge.

### *3.6. Precise Nutrition and Health Management*

The development of big data provides technical support for the processing of massive data, and scientific guidance for human nutrition and health management. In the past, people usually learned nutrition information from experts, books, and the Internet. However, there was a lack of accurate nutrition and health managemen<sup>t</sup> for individuals because of the difference in individual health conditions [67]. In an example of applying big data to the people's daily diet Teng et al. proposed to use a recipe recommendation algorithm to determine which food ingredients were necessary [68]. Grace et al. combined case-based reasoning and a deep learning algorithm to generate new recipes [69,70]. However, it may also generate "dark cuisine" due to the the uncertain factor of deep learning. Some scholars, like Freyne et al., focused on a diet therapy. They developed a personalized recipe recommendation system for obese people based on the suggestions from medical professionals and research on obese people [71]. In anoher instance, Yoshida et al. proposed a personalized recipe recommendation based on users' food preferences [72]. Zeevi et al. broke with traditional experience-based nutritional recommendations by using machine-learning algorithms to combine data (e.g., blood parameters, dietary habits, and gu<sup>t</sup> microbiota) to formulate personalized diets that optimize postprandial glucose levels and metabolites [73]. The combination of big data with Artificial Intelligence will provide a new approach for the research of precision nutrition.
