Citing NPAP#

If NPAP contributes to work that you publish, please cite it. Citations are what make sustained maintenance of research software possible.

Citing the software#

Every release of NPAP is described by the CITATION.cff file in the repository root. GitHub renders it as a “Cite this repository” button on the project page, which exports the entry in both APA and BibTeX format.

Please cite the specific version you used rather than the project in general, so that your results stay reproducible. You can find the version you are running with:

import npap

print(npap.__version__)

Citing the paper#

NPAP has been submitted to the Journal of Open Source Software; the review is open and in progress.

Note

Once the paper is accepted, the DOI and the definitive BibTeX entry will be published here and in CITATION.cff. Until then, please cite the software release as described above.

Citing the underlying methods#

NPAP builds directly on several scientific libraries. If your work depends on a particular component, consider citing it as well:

Component

Used for

Reference

NetworkX

Graph data structures and algorithms

Hagberg, Schult & Swart (2008)

scikit-learn

k-means, DBSCAN, hierarchical clustering

Pedregosa et al. (2011)

SciPy

Sparse linear algebra and distance metrics

Virtanen et al. (2020)

kmedoids

FasterPAM k-medoids implementation

Schubert & Rousseeuw (2021)

HDBSCAN

Density-based clustering

McInnes, Healy & Astels (2017)

If your networks come from a published dataset — for example the PyPSA-Eur network used in the examples — please cite that dataset separately, following its own citation instructions.