Share a ydata-profiling or Sweetviz report as a link
ydata-profiling (formerly pandas-profiling) and Sweetviz save a dataset report as one HTML file. Upload that file and send a link, so the people who need the report can open it in a browser without Python or Jupyter.
A profiling report is the quickest way to show someone what a dataset contains: missing values, distributions, correlations and duplicate rows. It is generated in Python, but the people who need to read it often do not run Python.
Both libraries write the report to a single HTML file. Emailing that file tends to fail: many mail apps open attachments as text or block them, and a report on a large dataset can pass attachment size limits. A link avoids both.
Save the report to an HTML file
ydata-profiling:
- ✓from ydata_profiling import ProfileReport
- ✓profile = ProfileReport(df, title="Orders data")
- ✓profile.to_file("orders-report.html")
Sweetviz: report = sweetviz.analyze(df), then report.show_html("orders-report.html", open_browser=False).
A table on its own needs no library: df.to_html("table.html") writes the DataFrame as an HTML table.
Reports on wide or very large datasets can run to tens of megabytes. ProfileReport(df, minimal=True) skips the most expensive sections, and profiling a sample with df.sample(n=100_000) keeps the file smaller.
Upload it and send the link
Drag the .html file onto the FileShare Pro dashboard and copy the link. The report opens as a web page in any browser, including on a phone, and viewers need no account. Tabs, collapsible sections and toggles work as they did locally.
If parts of the page look unstyled once hosted, the file may point to a local folder for its CSS or scripts. Making an HTML file self-contained covers how to find and fix those references.
Collect questions on the report
Turn on comments and readers can leave a note on the page after typing a name, which keeps questions about a column or an outlier next to the report. The view count shows whether the people you sent it to have opened it.
When the data is refreshed, generate the report again and upload it over the old one. The link stays the same: updating a shared link without changing it.
A dataset that should not be public belongs behind a password, which is on Pro: password-protecting an HTML page. Notebooks and Plotly charts follow a similar path: see sharing a Jupyter notebook as a link.
Frequently asked
Save it with profile.to_file("report.html"), upload the HTML file to a host that serves it as a web page, and send the link. It opens in any browser.
Yes. The package was renamed to ydata-profiling. Import ProfileReport from ydata_profiling; to_file() works the same way.
Upload it and send a link instead. The free plan accepts files up to 100 MB. To make the report smaller, use minimal=True or profile a sample of the rows.
It contains summaries and, by default, sample rows from the dataset. Anyone with the link can open it, so remove sensitive columns before profiling or use a password-protected link on Pro.
Yes. The free plan hosts up to 50 files of up to 100 MB each, with view counts and comments included.
Drop the file, copy the link — live in seconds, free plan included.
Start sharing free