Explore and label on a map of raw data.

Get enough to feed your model in no time.

hover speeds up data labeling through embedding + visualization + callbacks.

• You just need raw data and an embedding to get a map.
• Explore the map to find "zones" that are easy and those that are tricky.
• Start the conquest of your data by coloring the zones through wisdom!

🚀 Live Demos

With code

• edit & run code in your browser to get a labeling interface, with guides along the way.

Without code

• go directly to a labeling interface, pre-coded and hosted on Binder.

✨ Features (click bullets to show GIF)

It's fast because it labels data in bulk.

🔭 A semantic scatter plot of your data for labeling, equipped with

• Tooltip for each point on mouse hover

  
• Table view for inspecting all selected points

  
• Toggle buttons that clearly distinguish data subsets

  
• Search widgets for ad-hoc data highlight

  

It's accurate because multiple angles work together.

🔬 Supplementary views to use in conjunction with the annotator, including

• Finder: filter data by search criteria

  
• Softlabel: active learning by in-the-loop model prediction score

  
• Snorkel: custom functions for labeling and filtering

  

It's flexible (and fun!) because the process never gets old.

🧰 Additional tools and options that allow you to

• Go to higher dimensions (3D? 4D?) and choose your xy-axes

  
• Consecutively select across areas, dimensions, and views

  
• Kick outliers and fix mistakes

  

💡 Click on the bullet points above to view GIFs!

📦 Install

Python: 3.8+

OS: Linux & Mac & Windows

PyPI (for all releases): pip install hover

Conda-forge (for 0.6.0 and above): conda install -c conda-forge hover

For Windows users, we recommend Windows Subsystem for Linux.

• On Windows itself you will need C++ build tools for dependencies.

📖 Resources

• Tutorials
• Binder repo
• Changelog
• Documentation

🎏 Announcements

• Jan 21, 2023 version 0.8.0 is now available. Check out the changelog for details 🥳.

🔔 Remarks

Shoutouts

• Thanks to Bokeh because hover would not exist without linked plots and callbacks, or be nearly as good without embeddable server apps.
• Thanks to Philip Vollet for sharing hover with the community even when it was really green.

Contributing

• All feedbacks are welcome, especially what you find lacking and want it fixed!
• ./requirements-dev.txt lists required packages for development.
• Pull requests are advised to use a superset of the pre-commit hooks listed in .pre-commit-config.yaml.

Citation

If you have found hover useful to your work, please let us know 🤗

@misc{hover,
  title={{hover}: label data at scale},
  url={https://github.com/phurwicz/hover},
  note={Open software from https://github.com/phurwicz/hover},
  author={
    Pavel Hurwicz and
    Haochuan Wei},
  year={2021},
}