数据集与探索性数据分析

1import numpy as np
2import pandas as pd
3import matplotlib.pyplot as plt
4import seaborn as sns
5sns.set(font_scale=1.3) # setting font size for the whole notebook
6sns.set_style("whitegrid") # setting the style

1df = pd.read_csv('Coded_Data.csv')
2print(df.head(2))

1df = pd.read_csv('Coded_Data.csv',index_col=0)
2print(df.head(2))

1print(df.info())

1print(df.describe())

1means_stds = df.describe().T[['mean', 'std']]
2print(means_stds)

1means_stds.plot(kind='bar',figsize=(12,6));
2plt.yscale('log')
3plt.xlabel('features')
4plt.ylabel('mean (log scale)');

1sns.pairplot(df, hue='Result')

1print(df.Result.value_counts())