Pita and pizza restaurants and menus in Ankara

This study investigates whether the ingredients listed on restaurant menus can provide insights into a city's socioeconomic status. Using data from an online food delivery system, the study compares menu items with local education rates and rental prices. A machine learning model is developed to predict menu prices based on ingredients and socioeconomic factors. An efficiency metric is proposed to cluster restaurants to address autocorrelation, comparing ingredient averages to socioeconomic indicators. The analysis focuses on hundreds of menus, specifically examining pizza and Turkish pita in Ankara, Türkiye. The results indicate that including nearby rental prices significantly improves the accuracy of predicting menu prices, especially for pizza. The study also notes that wealthier areas tend to feature menus with more unique or expensive ingredients, particularly in the case of pizza, aligning with previous research on eating habits and income levels. Key contributions of this research include a comprehensive examination of restaurant menus, insights into how menus vary based on location and cuisine, and the development of Turkish-English word lists for pita and pizza menu items. Our datasets are also shared. This methodology aids in understanding local taste preferences and provides valuable information for strategic decisions regarding restaurant location and menu planning.

本研究旨在探讨餐馆菜单上列出的食材是否能够为城市的经济社会地位提供洞见。通过利用在线食品配送系统的数据，研究将菜单项目与当地的教育率和租金价格进行比较。本研究开发了一种机器学习模型，用以根据食材和经济社会因素预测菜单价格。此外，提出了一种效率指标，用于对餐馆进行聚类以解决自相关性问题，即将食材的平均值与经济社会指标进行对比。分析集中于数百份菜单，特别考察了安卡拉地区的比萨饼和土耳其烤饼。结果显示，包含附近的租金价格显著提高了预测菜单价格准确性，尤其是在比萨饼方面。研究还指出，较富裕的地区往往拥有更多独特或昂贵的食材菜单，特别是在比萨饼的情况下，这与以往关于饮食习惯和收入水平的研究结果相吻合。本研究的核心贡献包括对餐馆菜单的全面考察、对菜单如何根据地理位置和菜系而变化的见解，以及为比萨饼和土耳其烤饼菜单项目开发土耳其-英语词汇表。我们的数据集也予以共享。此方法论有助于理解当地口味偏好，并为关于餐馆位置和菜单规划的策略决策提供宝贵信息。