Hagelslag

An object-based severe storm forecasting system that utilizing image processing and machine learning tools to derive calibrated probabilities of severe hazards from convection-allowing numerical weather prediction model output.
https://github.com/djgagne/hagelslag

geojson hail hrrr machine-learning mrms netcdf performance performance-diagram python reliability segmentation storms tracking verification weather zarr

Last synced: about 2 hours ago
JSON representation

Repository metadata

Hagelslag supports segmentation and tracking of weather fields and scalable verification, including performance diagrams and reliability diagrams.

• Host: GitHub
• URL: https://github.com/djgagne/hagelslag
• Owner: djgagne
• License: mit
• Created: 2015-06-16T20:48:43.000Z (almost 9 years ago)
• Default Branch: master
• Last Pushed: 2024-02-10T00:50:36.000Z (3 months ago)
• Last Synced: 2024-04-25T05:06:11.009Z (16 days ago)
• Topics: geojson, hail, hrrr, machine-learning, mrms, netcdf, performance, performance-diagram, python, reliability, segmentation, storms, tracking, verification, weather, zarr
• Language: Jupyter Notebook
• Homepage:
• Size: 91.3 MB
• Stars: 67
• Watchers: 16
• Forks: 27
• Open Issues: 6
• Metadata Files:
  • Readme: README.md
  • License: LICENSE.txt

README

        
# Hagelslag

## Storm tracking, machine learning, and probabilistic evaluation

[![NSF-1261776](https://img.shields.io/badge/NSF-1261776-blue)](https://www.nsf.gov/awardsearch/showAward?AWD_ID=1261776&HistoricalAwards=false)

Hagelslag is an object-based severe storm forecasting system that utilizing image processing and machine learning tools

to derive calibrated probabilities of severe hazards from convection-allowing numerical weather prediction model output.

The package contains modules for storm identification and tracking, spatio-temporal data extraction, and 

machine learning model training to predict hazard intensity as well as space and time translations.

### Citation

If you employ hagelslag in your research, please acknowledge its use with the following citations:

    Gagne, D. J., A. McGovern, S. E. Haupt, R. A. Sobash, J. K. Williams, M. Xue, 2017: Storm-Based Probabilistic Hail

    Forecasting with Machine Learning Applied to Convection-Allowing Ensembles, Wea. Forecasting, 32, 1819-1840. 

    https://doi.org/10.1175/WAF-D-17-0010.1. 

    

    Gagne II, D. J., A. McGovern, N. Snook, R. Sobash, J. Labriola, J. K. Williams, S. E. Haupt, and M. Xue, 2016: 

    Hagelslag: Scalable object-based severe weather analysis and forecasting. Proceedings of the Sixth Symposium on 

    Advances in Modeling and Analysis Using Python, New Orleans, LA, Amer. Meteor. Soc., 447.

If you discover any issues, please post them to the Github issue tracker page. Questions and comments should be sent to

djgagne at ou dot edu.

### Requirements

Hagelslag is compatible with Python 3.6 or newer. Hagelslag is easiest to install with the help of the [Miniconda 

Python Distribution](https://docs.conda.io/en/latest/miniconda.html), but it should work with other

Python setups as well. Hagelslag requires the following packages and recommends the following versions:

* numpy >= 1.10

* scipy >= 0.15

* matplotlib >= 1.4

* scikit-learn >= 0.16

* pandas >= 0.15

* arrow >= 0.8.0

* pyproj

* netCDF4-python

* xarray

* jupyter

* ncepgrib2

* pygrib

* cython

* pip

* sphinx

* mock

Install dependencies with the following commands:

```

git clone https://github.com/djgagne/hagelslag.git

cd ~/hagelslag

conda env create -f environment.yml

conda activate hagelslag

```

### Installation

Install the latest version of hagelslag with the following command from the top-level hagelslag directory (where setup.py

is):

`pip install .`

Hagelslag will install the libraries in site-packages and will also install 3 applications into the `bin` directory

of your Python installation.

### Use

A Jupyter notebook is located in the demos directory that showcases the functionality of the package. For larger scale 

use, 3 scripts are provided in the bin directory. 

* `hsdata` performs object tracking and matching as well as data processing.

* `hsfore` trains and applies machine learning models.

* `hseval` performs forecast verification.

All scripts take input from a config file. The config file should be valid Python code and contain a dictionary called

config. Custom machine learning models and parameters should be contained within the config files. Examples of them can

be found in the config directory.

### Documentation

API Documentation is available [here](http://hagelslag.readthedocs.io/en/latest/).

      

---

Owner metadata

• Name: David John Gagne
• Login: djgagne
• Email:
• Kind: user
• Description: Machine Learning Scientist at the National Center for Atmospheric Research.
• Website: https://djgagne.github.io
• Location: Boulder, CO
• Twitter:
• Company: National Center for Atmospheric Research
• Icon url: https://avatars.githubusercontent.com/u/874553?u=f24a96dac586f5631cc30383eb58de2cca75f1cc&v=4
• Repositories: 16
• Last ynced at: 2023-03-24T05:17:03.391Z
• Profile URL: https://github.com/djgagne

---

GitHub Events

Total

• Create event: 10
• Commit comment event: 1
• Release event: 5
• Delete event: 4
• Member event: 2
• Pull request event: 86
• Fork event: 27
• Issues event: 6
• Watch event: 66
• Issue comment event: 36
• Public event: 1
• Push event: 293
• Pull request review event: 21
• Pull request review comment event: 17

Last Year

• Create event: 2
• Delete event: 1
• Fork event: 1
• Issue comment event: 2
• Issues event: 2
• Pull request event: 7
• Pull request review event: 2
• Push event: 14
• Release event: 1
• Watch event: 7

---

Committers metadata

Last synced: 1 day ago

Name	Email	Commits
David John Gagne	d****e@o****u	362
ahijevyc	a****c@u****u	73
David John Gagne	d****e@g****m	64
Amanda Burke	a****e@n****t	35
Amanda Burke	3****e	31
Ubuntu	t****o@e****t	10
Charlie Becker	c****r@g****m	9
Charlie Becker	3****r	6
Amanda Burke	a****e@l****u	5
Amanda Burke	a****e@l****u	4
Amanda Burke	a****e@l****u	3
Amanda Burke	a****e@s****u	3
Kate Avery	k****y@o****u	3
Nathan Wendt	n****t@o****u	3
Amanda Burke	a****e@d****u	2
Amanda Burke	a****e@l****u	2
Thomas Martin	3****o	1
David Gagne	d****e@s****u	1
Luke Madaus	m****e@g****m	1
Maria J. Molina	h****a@g****m	1

Committer domains:

• ou.edu: 3
• stratus.caps.ou.edu: 2
• login4.stampede2.tacc.utexas.edu: 1
• d-ip-10-197-1-103.nwc.nor.ou.edu: 1
• login2.stampede2.tacc.utexas.edu: 1
• login1.stampede2.tacc.utexas.edu: 1
• login3.stampede2.tacc.utexas.edu: 1
• edmonton.l44ohl0bwfwulknb5fapusimwc.jx.internal.cloudapp.net: 1
• nimbus.capsint: 1
• ucar.edu: 1

---

Issue and Pull Request metadata

Last synced: 1 day ago

More stats: https://issues.ecosyste.ms/repositories/lookup?url=https://github.com/djgagne/hagelslag

---

---

---

Package metadata

• Total packages: 1
• Total downloads:
  • pypi: 152 last-month
• Total dependent packages: 0
• Total dependent repositories: 2
• Total versions: 5
• Total maintainers: 1

---

Dependencies

Score: 12.329616831621628