Example

First, load packages and an example dataset:

import numpy as np
import bayspline as bsl

example_file = bsl.get_example_data('tierney2016-p178-15p.csv')
d = np.genfromtxt(example_file, delimiter=',', names=True)

This dataset (from Tierney et al. 2015) has three columns giving core depth (cm), sediment age (calendar years BP), and UK’37.

We can predict sea-surface temperatures (SST) from UK’37 with bsl.predict_sst():

prediction = bsl.predict_sst(d['uk37'], prior_std=10)

To see actual numbers from the prediction, directly parse prediction.ensemble or use prediction.percentile() to get the 5%, 50% and 95% percentiles.

You can also plot your prediction with bsl.predictplot() or bsl.densityplot().

Alternatively, we can make inferences about UK’37 from SST with bsl.predict_uk():

sst = np.arange(1, 25)
prediction = bsl.predict_uk(sst)

Installation

Install baysplinepy in conda with:

$ conda install baysplinepy -c sbmalev

To install with pip, run:

$ pip install baysplinepy

Unfortunately, baysplinepy is not compatible with Python 2.

Support and development

• Please feel free to report bugs and issues or view the source code on GitHub (https://github.com/brews/baysplinepy).

License

baysplinepy is available under the Open Source GPLv3 (https://www.gnu.org/licenses).