The Covid-19 CODO Development Process: an Agile Approach to Knowledge Graph Development

EasyChair Preprint 6802

Abstract

The CODO ontology was designed to capture data about the Covid-19 pandemic. The goal of the ontology was to collect epidemiological data about the pandemic so that medical professionals could perform contact tracing and answer questions about infection paths based on information about relations between patients, geography, time, etc. We took information from various spreadsheets and integrated it into one consistent knowledge graph that could be queried with SPARQL and visualized with the Gruff tool in AllegroGraph. The ontology is published on Bioportal and has been used by two projects to date. This paper describes the process used to design the initial ontology and to develop transformations to in-corporate data from the Indian government about the pandemic. We went from an ontology to a large knowledge graph with approximately 5M triples in a few months. Our experience demonstrates some common principles that apply to the process of scaling up from an ontology model to a knowledge graph with large amounts of real-world data.

Keyphrases: Agile Methods, Knowledge Graph, OWL, Ontology, SPARQL, Software Development Life Cycle, Web Ontology Language, transformations

@booklet{EasyChair:6802,
  author    = {Michael DeBellis and Biswanath Dutta},
  title     = {The Covid-19 CODO Development Process: an Agile Approach to Knowledge Graph Development},
  howpublished = {EasyChair Preprint 6802},
  year      = {EasyChair, 2021}}