SciPy Cheat Sheet

**from** *scipy.stats* **import** *ttest_1samp*

tstat, pval = **ttest_1samp***(example_distribution, expected_mean)*
-> *Generates two outputs* 
-> **tstat:***t-statistic*
-> **pval:*** p-value*

**from** *scipy.stats* **import** *ttest_ind*

tstat, pval = ttest_ind(data1, data2)

**from** *scipy.stats* **import** *f_oneway*

fstat, pval = f_oneway(data1, data2, data3)

**from** *statsmodels.stats.multicomp* **import** *pairwise_tukeyhsd*

# *All Data has to be unioned to one List*
movie_scores = np.**concatenate**([drama_scores, comedy_scores, documentary_scores])
labels = ['drama'] \* len(drama_scores) + ['comedy'] \* len(comedy_scores) + ['documentary'] \* len(documentary_scores)

**tukey_results = pairwise_tukeyhsd(movie_scores, labels, 0.05)**
-> *0.05 represents the significance level*

**from** *scipy.stats* **import** *binom_test*

pval = binom_test(*successes*, n, p)

*successes: * Number observed successes
n: Number of Trials
p: Expected probability for success

-> if pval < 0.05 we can reject the Null Hypothesis

**from** *scipy.stats* **import** *chi2_contingency*

_, pval, _, _ = chi2_contingency(iron_contingency_table)