Untitled

import pandas as pd
import os
import matplotlib.pyplot as plt
import scipy as scp
from scipy.stats import spearmanr
import seaborn as sns
import numpy as np
import matplotlib
matplotlib.rcParams['figure.figsize'] = [6, 6]
######
df = pd.read_csv('Austrian Survey on Privacy.csv')
df_KTP = df.iloc[:, [0,6,7,8,9,10] + list(range(145, 148)) + [169, 171, 177]]
df_PP = df.iloc[:, [0,6,7,8,9,10,300,306,314,315,316,317,318,319,326,327,328,329,330,331,332]]
df_KTP['knowledge_score'] = df_KTP.iloc[:,7:13].sum(axis=1, skipna=True)
df_PP['trust_score_authorities'] = df_PP.iloc[:,8:15].sum(axis=1, skipna=True)
df_PP['trust_score_big_tech'] = df_PP.iloc[:,15:20].sum(axis=1, skipna=True)
df_PP['trust_score_authorities'] = df_PP['trust_score_authorities'].divide(6, fill_value=np.nan)
df_PP['trust_score_big_tech'] = df_PP['trust_score_big_tech'].divide(5, fill_value=np.nan)
df_PP['trust_score'] = df_PP.iloc[:,[14,20,21,22]].sum(axis=1, skipna=True)
df_PP['trust_score'] = df_PP['trust_score'].replace(0, np.nan)
df_KTP['knowledge_score'] = df_KTP['knowledge_score'].replace(0, np.nan)
df_PP['trust_score_authorities'] = df_PP['trust_score_authorities'].replace(0, np.nan)
df_PP['trust_score_big_tech'] = df_PP['trust_score_big_tech'].replace(0, np.nan)
######
df_PP['Please choose your age group'] = df_PP['Please choose your age group'].replace('Younger than 18', 1)
df_PP['Please choose your age group'] = df_PP['Please choose your age group'].replace('18-24', 1)
df_PP['Please choose your age group'] = df_PP['Please choose your age group'].replace('25-34', 2)
df_PP['Please choose your age group'] = df_PP['Please choose your age group'].replace('35-44', 3)
df_PP['Please choose your age group'] = df_PP['Please choose your age group'].replace('45-54', 4)
df_PP['Please choose your age group'] = df_PP['Please choose your age group'].replace('55-64', 5)
df_PP['Please choose your age group'] = df_PP['Please choose your age group'].replace('65-74', 6)
df_PP['Please choose your age group'] = df_PP['Please choose your age group'].replace('Older than 75', 6)

df_KTP['Please choose your age group'] = df_KTP['Please choose your age group'].replace('Younger than 18', 1)
df_KTP['Please choose your age group'] = df_KTP['Please choose your age group'].replace('18-24', 1)
df_KTP['Please choose your age group'] = df_KTP['Please choose your age group'].replace('25-34', 2)
df_KTP['Please choose your age group'] = df_KTP['Please choose your age group'].replace('35-44', 3)
df_KTP['Please choose your age group'] = df_KTP['Please choose your age group'].replace('45-54', 4)
df_KTP['Please choose your age group'] = df_KTP['Please choose your age group'].replace('55-64', 5)
df_KTP['Please choose your age group'] = df_KTP['Please choose your age group'].replace('65-74', 6)
df_KTP['Please choose your age group'] = df_KTP['Please choose your age group'].replace('Older than 75', 6)
######
sns.regplot(data=df_KTP, x='Please choose your age group', y='knowledge_score', line_kws={'color': 'red'})
sns.violinplot(data=df_KTP, y='How confident are you with using digital technologies?', x='Please choose your age group')
corr, pval = spearmanr(df_PP['Please choose your age group'], df_PP['trust_score'], nan_policy='omit')