Customer's Need
The client, a company operating in the Organized Retail sector (GDO) in South American markets, needed to effectively monitor online product categories. The goal was to obtain a detailed and real-time analysis of the range of products available on various online GDO sites. This monitoring had to include a wide range of product categories to provide the client with a comprehensive overview of the market.
Feature 1
Lorem ipsum dolor sit amet, consectetur adipiscing elit.
Feature 2
Lorem ipsum dolor sit amet, consectetur adipiscing elit.
Feature 5
Lorem ipsum dolor sit amet, consectetur adipiscing elit.
Implemented Solution
Crawling and Scraping
We developed a daily crawling and scraping system of product pages for entire categories present on all online GDO sites in South American markets. This allowed us to efficiently and systematically collect a vast amount of data regarding available products on the market.
Acquisition of Technical Sheet Properties
During the scraping process, we accurately and completely extracted the properties and technical specifications of individual products from their respective technical sheets. This allowed us to create a database rich in detailed information about available products, essential for data analysis and visualization.
Faceted Filtering Dashboard
We developed a customized dashboard that allows the client to filter and view data based on various categories, properties, and keywords. This enables the client to quickly obtain relevant information for their specific market analysis and monitoring needs.
Machine Learning System
We implemented a machine learning system to automatically determine the brand and model of products when not explicitly indicated in the collected information. This significantly improved the accuracy and completeness of the collected data, eliminating the need for manual corrective interventions.
Machine Learning for Product Similarity Determination
Finally, we developed a machine learning algorithm to determine the similarity of products using images acquired during crawling. This allowed us to identify and group similar products, further facilitating data analysis and interpretation for the client.