Web Scraping in Supermarkets for Monitoring Basic Food Basket Prices

Abstract

Tracking the prices of the basic food basket is fundamental to ensure food security and economic stability of a country. For this purpose, a set of Python scripts were developed to collect price data from five supermarket websites. For this purpose, the Selenium browser automator was used, which together with libraries such as Pandas and Numpy, allowed the consolidation of the information corresponding to the name of the products, their descriptions, their prices and the dates of information capture for a period of time of 100 days. Finally, a series of dashboards were created in Power BI in order to know the experience of a group of users with the information presented. The results of the concept test with users identified aspects of presentation, search, comprehension, inference and decision making based on the visualizations presented. Regarding this last aspect, users pointed out that the absence of information related to the price per gram or per milliliter for the products does not allow them to make comparisons that lead them to clearly differentiate the prices of these products. This is an element that will be addressed in a future paper.