Pandas Profiling Versions Save

Dependency policy

The current dependency policy is suboptimal. Pinning the dependencies is great for reproducibility (high guarantee to work), but on the downside requires frequent maintenance and introduces compatibility issues with other packages. Therefore, we are moving away from pinning dependencies and instead specify a minimum version.

Pandas v1

Early releases of pandas v1 demonstrated many regressions that broke functionality (as acknowledged by the authors here). At this point, pandas is more stable and we notice high demand for compatibility. We move on to support pandas' latest versions. To ensure compatibility with both versions, we have extended the test matrix to test against both pandas 0.x.y and 1.x.y.

Python 3.6+ features

Python 3.6 introduces ordered dicts and f-strings, which we now rely on. This means that from pandas-profiling 2.6, you should minimally run Python 3.6. For users that for some reason cannot update, you can use pandas-profiling 2.5.0, but you unfortunately won't benefit from updates or maintenance.

Extended continuous integration

Starting from this release, we use Github Actions and Travis CI combined to increase maintainability. Travis CI handles the testing, Github Actions automates part of the development process by running black and building the docs.

• Progress bar added (#224)
• Character analysis for Text/NLP (#278)
• Themes: configuration and demo's (Orange, Dark)
• Tutorial on modifying the report's structure (#362; #281, #259, #253, #234). This jupyter notebook also demonstrates how to use the Kaggle api together with pandas-profiling.
• Toggle descriptions at correlations.

Deprecation:

• This is the last version to support Python 3.5.

Stability:

• The order of columns changed when sort="None" (#377, fixed).
• Pandas v1.0.X is not yet supported (#367, #366, #363, #353, pinned pandas to < 1)
• Improved mixed type detection (#351)
• Refactor of report structures.
• Correlations are more stable (e.g. Phi_k color scale now from 0-1, rows and columns with NaN values are dropped, #329).
• Distinct counts exclude NaNs.
• Fixed alerts in notebooks.

Other improvements:

• Warnings are now sorted.
• Links to Binder and Google Colab are added for notebooks (#349)
• The overview section is tabbed.

• Commit for pandas-profiling v2.5.0

• Progress bar added (#224)
• Character analysis for Text/NLP (#278)
• Themes: configuration and demo's (Orange, Dark)
• Tutorial on modifying the report's structure (#362; #281, #259, #253, #234). This jupyter notebook also demonstrates how to use the Kaggle api together with pandas-profiling.
• Toggle descriptions at correlations.

Deprecation:

• This is the last version to support Python 3.5.

Stability:

• The order of columns changed when sort="None" (#377, fixed).
• Pandas v1.0.X is not yet supported (#367, #366, #363, #353, pinned pandas to < 1)
• Improved mixed type detection (#351)
• Refactor of report structures.
• Correlations are more stable (e.g. Phi_k color scale now from 0-1, rows and columns with NaN values are dropped, #329).
• Distinct counts exclude NaNs.
• Fixed alerts in notebooks.

Other improvements:

• Warnings are now sorted.
• Links to Binder and Google Colab are added for notebooks (#349)
• The overview section is tabbed.

The v2.4.0 release decouples the data structure of reports from the actual rendering. It's now much simpler to change the user interface, whether the user is in a jupyter notebook, webpage, native application or just wants a json view of the data.

We are also proud to announce that we are accepted for the GitHub Sponsor programme. You are cordially invited to support me through this programme, because you want to see me continue working on this project and to boost community funding, GitHub will match your contribution!

Other improvements:

• extended configuration with better defaults, including minimal mode for big data (#258, #310)
• more example datasets
• rejection of highly correlated variables is generalized (#284, #299)
• many structural and stability improvements (#254, #274, #239)

Special thanks to @marco-cardoso @ajupton @lvwerra @gliptak @neomatrix369 for their contributions.

• (Experimental) Support for "path" type
• Fix numeric precision (#225)
• Force labels in missing values diagram for large number of columns (#222)
• Add pull request template
• Add Census Dataset from the UCI ML Repository

Thanks @bensdm and @huaiweicheng for your valuable contributions to this version!

New release introducing variable size binning (via astropy), PyCharm integration and various fixes and optimizations.

• Added Variable bin sizing via Bayesian Boxing (feature request [#216])
• PyCharm integration, console attempts to detect file type.
• Fixed bug [#215].
• Updated the missingno package to 0.4.2, fixing the font size in the bar diagram.
• Various optimizations

Thanks to: @Utsav37 @mansenfranzen @jakevdp

Fix [#211] and README

• Fix of [#206]
• Improve code maintainability of the view (HTML templates, notebook)
• Fix bug in dendrogram sizing

The pandas-profiling release version 2.1.0 includes:

• Correlations: correlation calculations are now more fault tolerant ([#51] and [#197]), correlation names in the report are clarified.
• Jupyter Notebook: rendering a profiling report is done inside the srcdoc attribute (which fixes [#199]), a full-width option is added and the column layout is improved.
• User experience: The table styling and sample section formatting is improved.
• Warnings: detection added for categorical variable that is suspected to be of the datetime type.
• Documentation and community:
  • The Contribution page helps users that want to contribute.
  • Typo's fixed [#195], Thank you @abhilashshakti
  • Added more examples.
• Other bugfixes and improvements:
  • Add version information to console interface.
  • Fix: Remove one-time used logger [#202]
  • Fix: Dealing with string indices [#200]

Contributors: @abhilashshakti @adamrossnelson @manycoding @InsciteAnalytics

Bugfix on version structure for 2.0.2.

Revised version structure, fixed recursion preventing installation of dependencies ([#184]).

The setup.py file used to include utils from the package prior to installation. This causes errors when the dependencies are not yet present.