https://github.com/SamComber/spacv

Spatial cross-validation in Python.
https://github.com/SamComber/spacv

cross-validation data-science geographic-data-science machine-learning python scikit-learn scikitlearn-machine-learning sklearn spatial-data-science

Last synced: 10 days ago
JSON representation

Spatial cross-validation in Python.

• Host: GitHub
• URL: https://github.com/SamComber/spacv
• Owner: SamComber
• License: bsd-3-clause
• Created: 2020-06-12T08:07:15.000Z (over 4 years ago)
• Default Branch: master
• Last Pushed: 2024-05-19T11:17:27.000Z (6 months ago)
• Last Synced: 2024-10-11T08:21:24.467Z (25 days ago)
• Topics: cross-validation, data-science, geographic-data-science, machine-learning, python, scikit-learn, scikitlearn-machine-learning, sklearn, spatial-data-science
• Language: Jupyter Notebook
• Homepage:
• Size: 3.76 MB
• Stars: 43
• Watchers: 3
• Forks: 14
• Open Issues: 4
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

        
# `spacv`: spatial cross-validation in Python

`spacv` is a small Python 3 (3.6 and above) package for cross-validation of models

that assess generalization performance to datasets with spatial dependence. `spacv` provides

a familiar sklearn-like API to expose a suite of tools useful for points-based spatial prediction tasks.

See the notebook `spacv_guide.ipynb` for usage.







## Dependencies

* `numpy`

* `matplotlib`

* `pandas`

* `geopandas`

* `shapely`

* `scikit-learn`

* `scipy`

## Installation and usage

To install use pip:

    $ pip install spacv

Then build quick spatial cross-validation workflows with `sklearn` as:

```python

import spacv

import geopandas as gpd

from sklearn.model_selection import cross_val_score

from sklearn.svm import SVC

df = gpd.read_file('data/baltim.geojson')

XYs = df['geometry']

X = df[['NROOM', 'BMENT', 'NBATH', 'PRICE', 'LOTSZ', 'SQFT']]

y = df['PATIO']

# Build fold indices as a generator

skcv = spacv.SKCV(n_splits=4, buffer_radius=10).split(XYs)

svc = SVC()

cross_val_score(svc,       # Model 

                X,         # Features

                y,         # Labels

                cv = skcv) # Fold indices

```