frosted-tracks 1.1.0

Behavioral segmentation and clustering of trajectory data.

Frosted Tracks

Analysis of trajectory behavior using TICC and DBSCAN

Repository Structure

• src/: Source code

Best practice: prototype code in a notebook, then move it into src/frosted_tracks with proper docstrings and test cases when it's ready to share. Open a pull request to our GitHub repository if you'd like to integrate your work into the main trunk!

Python Environment

We recommend that you use Anaconda (https://www.anaconda.com) for your Python environment. If you do, there's an environment.yml file in this repository that you can use to set up your dependencies as follows:

conda env create -f environment.yml

License

See the file LICENSE in the root directory of the repository for details. We release this work under a 3-clause BSD license.

Changes

Version 1.0: Initial release. Not distributed to PyPI.

Version 1.1: Experimental cluster predictor disconnected. It was causing build errors when we tried to construct wheels. You must now supply the desired number of clusters when you call cluster_trajectories.