A missing repository file, a malformed notebook, and a valid notebook that GitHub cannot preview are different failures. Check them separately so a display problem does not turn into an unnecessary rewrite of the analysis.
Confirm the file and revision first
Check the repository, selected branch or commit, folder, and .ipynb filename. If a colleague sees an access error, confirm their access to that repository through your normal process. Changing notebook metadata cannot fix an inaccessible or incorrect address.
For a review, record the exact revision you inspected. A branch can change after someone sends its link. Compare the downloaded file with the intended revision before diagnosing differences in outputs.
Distinguish missing interaction from missing content
Suppose an illustrative capacity notebook shows a static chart but its custom JavaScript range control does nothing on GitHub. That result is consistent with GitHub's documented static renderer. It does not by itself show that the calculations failed.
If the reviewer needs that interaction, GitHub points to nbviewer for JavaScript rendering and a local notebook server for fully interactive notebook use. Confirm the notebook's requirements before choosing either. Do not send private repository contents to another service merely to investigate a preview.
The existing notebook sharing guide covers choosing a report or a runnable source deliverable. Here the immediate question is why the repository's viewer cannot show the expected content.
Check structure without executing notebook cells
Keep the downloaded original unchanged. In an environment that already has nbformat, save this short script beside a working copy named capacity-review.ipynb:
import nbformat
notebook = nbformat.read(
"capacity-review.ipynb", as_version=nbformat.NO_CONVERT
)
nbformat.validate(notebook)
print("Notebook structure is valid")
Reading with NO_CONVERT preserves the supplied format choice; validate checks it against the notebook schema. This script does not run the notebook's code cells. If it raises an error, retain the error details and the original file before attempting a specific repair. A successful structural check does not verify the analysis or its stored outputs. nbformat API
Try the documented local rendering fallback
Where nbconvert is available, run:
jupyter nbconvert --to html capacity-review.ipynb
GitHub documents this fallback for notebook rendering trouble. Open the generated HTML and compare the expected text, tables, and charts. This conversion command does not request cell execution; do not add --execute merely to get a preview of an unfamiliar notebook. nbconvert command options
If local conversion succeeds and GitHub's preview still fails for the same file, keep those observations together. If local conversion fails too, investigate its specific error. Neither outcome alone proves a GitHub outage or identifies one universal metadata fix.
Finish the review against the known revision. Keep a note of which viewer displayed which content, and distinguish a readable static chart from an analysis you actually reran.