spssmirror 2.1.0

A unified, self-contained statistical analysis library for Python -- an SPSS/R replacement.

SPSSMirror

A unified, self-contained statistical analysis library for Python -- an SPSS/R replacement.

What's new in 2.1.0

• Full pandas passthrough: any method not defined by SPSSMirror falls through automatically to the underlying pandas DataFrame -- mirror.head(), mirror.dtypes, mirror.groupby(...), mirror.fillna(...), all of pandas, completely unmodified.
• Flat API: every statistical method now lives directly on SPSSMirror (mirror.t_test_independent(...), mirror.linear_regression(...), mirror.cronbach_alpha(...)) -- no more .regression().linear(...) nesting required (the old grouped accessors like .regression() still work too, unchanged).
• Pandas-style output: every result renders as a real pandas Series or DataFrame (print(result) gives a clean aligned table; bare result in Jupyter renders as a grid).
• .hint(): mirror.hint() lists every available statistical method as a DataFrame; mirror.hint("regression") filters to one group.

from spssmirror import SPSSMirror

mirror = SPSSMirror().load_csv("survey.csv")
mirror.head()                                   # normal pandas
result = mirror.linear_regression("score ~ age + C(group)")
print(result)                                    # pandas-style table
mirror.hint("regression")                        # see what else is available

Installation

pip install spssmirror
pip install spssmirror[bayesian,timeseries,survival,multivariate]  # optional extras

License

MIT