Analytic Statistics

• Include common analytic statistical analyses

• Still, R is better for these tasks.

import numpy as np
import pandas as pd
import scipy.stats as stats

DEMO_DATA_ROOT = "../../../RepositoryData/data"

Kruskal Test

• Two independent sample means

hedges = pd.read_table(DEMO_DATA_ROOT + "/gries_sflwr/_inputfiles/04-1-2-1_hedges.csv")
hedges.head()

	CASE	HEDGES	SEX
0	1	17	F
1	2	17	F
2	3	17	F
3	4	17	F
4	5	16	F

u_statistic, p = stats.ks_2samp(hedges[hedges['SEX']=="M"]['HEDGES'],hedges[hedges['SEX']=="F"]['HEDGES'] )
print(u_statistic, '\n', p)

0.4666666666666667 
 0.0025300622362698397

Chi-square

data = np.array([[85, 65],
                 [100,147]])
data

array([[ 85,  65],
       [100, 147]])

V, p, df, expected = stats.chi2_contingency(data, correction=False)
print("Chi-square value = %1.2f, df = %1.2f, p = %1.2f"%(V, df, p))

Chi-square value = 9.82, df = 1.00, p = 0.00

McNear Test

• One dependent variable (categorical)

• dependent samples

data = pd.read_table(DEMO_DATA_ROOT + "/gries_sflwr/_inputfiles/04-1-2-3_accjudg.csv")
data.head()

	SENTENCE	BEFORE	AFTER
0	1	acceptable	acceptable
1	2	acceptable	acceptable
2	3	acceptable	acceptable
3	4	acceptable	acceptable
4	5	acceptable	acceptable

from statsmodels.sandbox.stats.runs import mcnemar

crosstab = pd.crosstab(data['BEFORE'],data['AFTER'])
x2, p = mcnemar(crosstab, correction=False)
print('Chi-square=%1.2f, p = %1.2f'%(x2, p))

Chi-square=13.00, p = 0.00

Independent t-test

vowels = pd.read_table(DEMO_DATA_ROOT + "/gries_sflwr/_inputfiles/04-3-2-1_f1-freq.csv")
vowels.head()

	CASE	HZ_F1	SEX
0	1	488.57	M
1	2	557.66	M
2	3	424.69	M
3	4	625.95	M
4	5	531.04	M

t, p = stats.ttest_ind(vowels[vowels['SEX']=='M']['HZ_F1'], vowels[vowels['SEX']=='F']['HZ_F1'])
print("t-score=%1.2f, p=%1.2f"%(t,p))

t-score=-2.44, p=0.02

One-way ANOVA

data = pd.read_table(DEMO_DATA_ROOT + "/gries_sflwr/_inputfiles/05-2_reactiontimes.csv")
data

	CASE	RT	FREQUENCY	FAMILIARITY	IMAGEABILITY	MEANINGFULNESS
0	almond	650.9947	0.693147	NaN	NaN	NaN
1	ant	589.4347	1.945910	med	hi	415.0
2	apple	523.0493	2.302585	hi	hi	451.0
3	apricot	642.3342	0.693147	lo	lo	NaN
4	asparagus	696.2092	0.693147	med	lo	442.0
...	...	...	...	...	...	...
72	tortoise	733.0323	1.386294	lo	lo	403.0
73	walnut	663.5908	2.484907	med	lo	468.0
74	wasp	725.7056	1.098612	NaN	NaN	NaN
75	whale	609.9745	0.000000	med	hi	474.0
76	woodpecker	686.3439	0.693147	NaN	NaN	NaN

77 rows × 6 columns

data = data.dropna()

from statsmodels.formula.api import ols
from statsmodels.stats.anova import anova_lm

model = ols('RT ~ FAMILIARITY', data).fit()
aov = anova_lm(model)
print(aov)

               df        sum_sq      mean_sq         F    PR(>F)
FAMILIARITY   2.0  16377.601187  8188.800594  4.645348  0.014654
Residual     45.0  79325.813100  1762.795847       NaN       NaN