esda 2.3.0
esda
ESDA is an open-source Python library for the exploratory analysis of spatial data. A subpackage of PySAL (Python Spatial Analysis Library), esda includes methods for global and local spatial autocorrelation analysis. See the tutorial for use cases.
Version 2.3.0 (2020-07-05)
Major Enhancements
- Highly performant multi-core, numba-based conditional permutation inference for local autocorrelation statistics (#116)
- Adbscan: an extension of the original DBSCAN algorithm that creates an ensemble of solutions generated by running DBSCAN on a random subset and "extending" the solution to the rest of the sample through nearest-neighbor regression (see Arribas-Bel, Garcia-Lopez & Viladecans-Marsal, 2020 for more details). (#120)
Detailed Changes
We closed a total of 38 issues (enhancements and bug fixes) through 14 pull requests, since our last release on 2020-02-01.
Issues Closed
- permutation inference performance using numba (#116)
- Adbscan fix (#120)
- BUG: fix join_count contingency table (#124)
- Preferred style format of new ESDA estimators (#118)
- G_Local returns unexpectedly low z_sim for what should be a hotspot (#112)
- [WIP]: initial commit of ripley in numpy-oriented style (#115)
- Tweak README.md (#117)
- Memory efficient conditional permutation for LISA (#113)
- ENH: More efficient Geary implementation with new test data (#114)
- Develop separate notebooks for functionalilty (#84)
- by_col is failing in test_by_col due to deprecation of from_item (#100)
- Join count tail-ness (#48)
- Noisy imports on 3.8 (#99)
- DeprecationWarning when running spatial smoothing (#111)
- ADBSCAN (#94)
- DOC: fixing link in geosilhouettes notebook (#110)
- Allow failures on 3.8 when pulling from git for testing. (#108)
- Docs for esda not reachable (#107)
- Tutorial (#106)
- Tutorial (#105)
- BUG: don't delete .nojekyll (#104)
- Docs (#103)
- REL: 2.2.1 bf release (#102)
- Fixes for pandas deprecation and 3.8 verbosity (#101)
Pull Requests
- permutation inference performance using numba (#116)
- Adbscan fix (#120)
- BUG: fix join_count contingency table (#124)
- Tweak README.md (#117)
- ENH: More efficient Geary implementation with new test data (#114)
- ADBSCAN (#94)
- DOC: fixing link in geosilhouettes notebook (#110)
- Allow failures on 3.8 when pulling from git for testing. (#108)
- Tutorial (#106)
- Tutorial (#105)
- BUG: don't delete .nojekyll (#104)
- Docs (#103)
- REL: 2.2.1 bf release (#102)
- Fixes for pandas deprecation and 3.8 verbosity (#101)
The following individuals contributed to this release:
- Serge Rey
- Levi John Wolf
- James Gaboardi
- Dani Arribas-Bel
- Jeffery Sauer
- Eli Knaap