https://github.com/cmudig/autoprofiler

Automatically profile dataframes in the Jupyter sidebar
https://github.com/cmudig/autoprofiler

jupyter pandas python

Last synced: 5 days ago
JSON representation

Automatically profile dataframes in the Jupyter sidebar

• Host: GitHub
• URL: https://github.com/cmudig/autoprofiler
• Owner: cmudig
• License: bsd-3-clause
• Created: 2022-03-24T00:51:58.000Z (over 2 years ago)
• Default Branch: main
• Last Pushed: 2024-01-21T18:16:39.000Z (10 months ago)
• Last Synced: 2024-11-07T19:12:14.812Z (7 days ago)
• Topics: jupyter, pandas, python
• Language: Svelte
• Homepage:
• Size: 24.3 MB
• Stars: 340
• Watchers: 5
• Forks: 11
• Open Issues: 28
• Metadata Files:
  • Readme: README.md
  • Contributing: CONTRIBUTING.md
  • License: LICENSE

Awesome Lists containing this project

README

        




    

        

    

     

        

    

     

        

    



Profile your [Pandas](https://pandas.pydata.org) Dataframes! Autoprofiler will automatically visualize your Pandas dataframes after every execution, no extra code necessary.

Autoprofiler allows you to spend less time specifying charts and more time interacting with your data by automatically showing you profiling information like:

-   Distribution of each column

-   Sample values

-   Summary statistics

## Updates profiles as your data updates

![screenshot of Autoprofiler](https://raw.githubusercontent.com/cmudig/AutoProfiler/main/.github/screenshots/profiler_sc.png)

Autoprofiler reads your current Jupyter notebook and produces profiles for the Pandas Dataframes in your memory as they change.

https://user-images.githubusercontent.com/13400543/199877605-ba50f9c8-87e5-46c9-8207-1c6496bb3b18.mov

## Install

To instally locally use pip and then open jupyter lab and the extension will be running.

```bash

pip install -U digautoprofiler

```

Please note, AutoProfiler only works in [JupyterLab](https://jupyter.org/install) with version >=3.x, < 4.0.0.

## Try it out

To try out Autoprofiler in a hosted notebook, use one of the options below

|                                                                                           Jupyter Lite                                                                                            |                                                                       Binder                                                                       |

| :-----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------: | :------------------------------------------------------------------------------------------------------------------------------------------------: |

| [![Lite](https://gist.githubusercontent.com/willeppy/35cdc20a3fc26e393ce76f1df35bcdfc/raw/a7fca1d0a2d62c2b49f60c0217dffbd0fe404471/lite-badge-launch-small.svg)](http://dig.cmu.edu/AutoProfiler) | [![Binder](https://mybinder.org/badge_logo.svg)](https://mybinder.org/v2/gh/cmudig/AutoProfiler/HEAD?labpath=examples%2FSF%20Housing%20Demo.ipynb) |

**Browser support:** AutoProfiler has been developed and tested with Chrome.

## Development Install

For development install instructions, see [CONTRIBUTING.md](CONTRIBUTING.md).

If you're having install issues, see [TROUBLESHOOTING.md](TROUBLESHOOTING.md).

## Acknowledgements

Big thanks to the Rill Data team! Much of our profiler UI code is adapted from [Rill Developer](https://github.com/rilldata/rill-developer).

## Citation

Please reference our [VIS'23 paper](https://arxiv.org/abs/2308.03964):

```bibtex

@article{epperson23autoprofiler,

  title={Dead or Alive: Continuous Data Profiling for Interactive Data Science},

  author={Will Epperson and Vaishnavi Goranla and Dominik Moritz and Adam Perer},

  journal={IEEE Transactions on Visualization and Computer Graphics},

  year={2023},

  url={https://arxiv.org/abs/2308.03964}

}

```

## Let us know what you think! 📢

We would love to hear your feedback on how you are using AutoProfiler! Please fill out [this form](https://forms.gle/V3ejpXxMcQXqYJG48) or email Will at [[email protected]](mailto:[email protected]).