Datathons and Software to Promote Reproducible Research

Area of Science:

• Medical Informatics
• Computational Biology
• Clinical Research

Background:

• Datathons promote interdisciplinary collaboration among clinicians, statisticians, and data scientists.
• These events have led to significant publications relevant to critical care research.
• Datathons offer a successful model for collaborative scientific endeavors.

Purpose of the Study:

• To introduce Chatto, an open-source software suite developed for datathon participants.
• To enhance teamwork and efficiency during collaborative data analysis.
• To improve the reproducibility of research findings.

Main Methods:

• Datathon participants formed multidisciplinary teams to address clinical questions.
• The Chatto software suite was provided to facilitate teamwork, data analysis, and query reformulation.
• Teams worked collaboratively over two days, with presentations to a panel of judges.

Main Results:

• Chatto significantly reduced research environment setup time from hours/days to minutes.
• The software proved effective in the datathon setting for collaborative analysis.
• Chatto continued to be a valuable tool for research beyond the event.

Conclusions:

• Chatto facilitates interdisciplinary teamwork via version control and discussion archiving.
• The software enhances research reproducibility by enabling post-publication analysis modification.
• Chatto aims to overcome challenges in collaborative data mining and improve research integrity.