We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Daniel Mietchen Computational reproducibility of Jupyter notebooks from biomedical publications |
Dive into the world of biomedical reproducibility with Daniel Mietchen's tool, "reproduce-me-kit," analyzing Jupyter notebooks for errors and suggestions.
- The speaker presents a tool called “reproduce-me-kit” that can analyze Jupyter notebooks from biomedical publications and provide insights on reproducibility.
- The tool can visualize certain aspects of reproducibility, such as execution time, errors, and dependencies.
- The speaker shares statistics on the number of articles in the corpus by journal, and notes that most notebooks are in Python.
- The tool can identify points in the publication process where errors or inconsistencies occur, such as duplicate cell numbers or incorrect dependencies.
- The speaker highlights the importance of standardization in reporting reproducibility and notes that current recommendations do not specify different types of stakeholders.
- The tool can provide recommendations for improving reproducibility, such as listing imported modules and using consistent naming conventions.
- The speaker discusses the importance of transparency in reporting and notes that many notebooks lack documentation on dependencies and libraries used.
- The tool can also identify potential issues with data availability and sharing, and provide recommendations for improving data sharing practices.
- The speaker invites the audience to play with the tool and provide feedback to improve its functionality.
- The talk also touches on the importance of computational reproducibility in scientific research and the potential benefits of using Jupyter notebooks as references.