1 of 8

Citizen Science

for

Text Annotation

2nd Network of BioThings Hackathon

Richard Good, Hannes Niedner, Andrew Su

2 of 8

The problem statement

  • The biomedical literature is massive and growing exponentially

​

  • Most biomedical knowledge is unstructured, making it difficult to access,query, compute.

3 of 8

Our goal

  • Structuring biomedical text is a potential Citizen Science application

​

    • Concept ID and semantic typing
    • Concept normalization
    • Relationship mapping

​

  • Hack together interfaces that test the limits of Citizen Science for text

4 of 8

Our approach

  • iPad app
  • Web-based application

5 of 8

The iPad app

  • Features
    • Concept ID and semantic typing
    • Concept normalization
    • Relationship mapping
  • Demo
  • Repo: to be posted soon

6 of 8

The Web app

  • BRAT: tool for professional annotators
    • Visual interface for annotating concepts & relations
    • Concept normalization features
    • Active user/developer community

7 of 8

BRAT for Citizen Science

​

8 of 8

Summary

Citizen Science success

Friday November 7

Functionality

Usability