Using Publicly-Available Data to Assess an Organization’s Research Scope

• Activity Number: 352 - Section on Statistical Consulting CPapers 1
• Type: Contributed
• Date/Time: Wednesday, August 5, 2020 : 10:00 AM to 2:00 PM
• Sponsor: Section on Statistical Consulting
• Author(s): Peter John De Chavez* and Jingting Hui
• Companies: PepsiCo and PepsiCo
• Keywords: Natural Language Processing; Multivariate Landscaping; Text Data

Abstract

Scientists from various companies constantly publish in peer-reviewed journals to showcase their depth of scientific expertise, their collaborations with public and private organizations, and obtain feedback from experts outside their organizations. Hence, there is a vast amount of information from public sources that a company can use to assess its research landscape vis-à-vis competitors. The landscape can then provide a data-driven approach to determine the areas where a company can potentially expand its research scope.

This paper assessed the research landscape of three companies by collecting relevant meta-information from more than 1000 publicly-available research abstracts. Natural Language Processing (NLP) was used to extract pertinent meta-information from the abstracts, and multivariate analysis was used to analyze the meta-information and visualize the research landscape of the three companies. The visualization showed the overlaps in the research agenda of the three companies, but more importantly, the data showed where there is a dearth of research for each company, giving each one of them valuable information on potential areas for expanding their research scope.