Jupyter Notebooks with Errors, 10,000 Examples

10,000 Jupyter notebooks, each containing at least one error, collected by JetBrains Research. The dataset includes notebook content and associated repository information, which may help in environment restoration. It was last updated on March 19, 2024.

Use Cases

• Training error detection models based on the presence of at least one error per notebook.
• Studying common error patterns in Jupyter notebooks based on the provided notebook content.
• Researching software development environments using the associated repository metadata.

Strengths

• Contains 10,000 individual Jupyter notebook examples.
• Each notebook is guaranteed to contain at least one error, providing a focused corpus for analysis.
• Includes repository metadata that may aid in environment reconstruction.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Last updated 2024-03-19 10:47:26; freshness should be verified.

Provenance

Source

JetBrains-Research

Collection Method

Likely gathered from public repositories containing Jupyter notebooks.

Freshness

2024-03-19 10:47:26