BOTTOM-UP METHODS FOR E-COMMERCE PRODUCT MAP ANALYSIS FROM A LARGE GROUP OF SMALL AND MEDIUM-SIZED WEBSITES

Abstract

Academics and industry professionals have shown considerable interest in the study of product 
maps in e-commerce because of the insights they give regarding the relationship between items, 
including their complementarity and rivalry. Existing research, however, has mostly relied on the 
input of big manufacturers and merchants and data collected from these authoritative sources. In 
this article, we use a crowdsourced bottom-up methodology, utilising SME e-commerce sites as 
the primary sources of information. This method permits distributed data processing and the 
synthesis of viewpoints and information from several distinct sources. In order to quantify 
product similarities and construct a product map, a graph term frequency-inverse document 
frequency approach is provided. Using information on more than 90,000 goods from 52 SME 
sites, a hierarchical community structure was discovered using this technique. The findings 
revealed a localised distribution pattern for goods produced on the same location. Our research 
may help online marketplaces increase their product variety via more calculated pricing and 
inventory management.