Kaggle: Human Resources Analytics

Introduce

这是新年定下的目标之一 —— Kaggle案例的学习
这里，数据即来自Human Resources Analytics， 参考这里学习数据的分析.

Overview

这个kernel里面呢，大部分还是比较常规的EDA，可视化和简单算法的应用，但是也有要学习的地方，这里在代码里面都有声明。

[学习heatmap使用]
[学习plt.barh使用]
[学习如下reindex的步骤]
[学习iloc使用]
[学习使用sklearn进行标准化]
[掌握手动计算协方差]
[学习使用Numpy简单的线性代数的运算]
[学习使用sklearn做PCA]

Code

#!/usr/bin/env python3
# -*- coding: utf-8 -*-
"""
Created on Sat Feb 25 07:48:50 2017

@author: shen
"""

# 人力资源——离职分析
# 采用PCA
# 参考https://www.kaggle.com/nirajvermafcb/d/ludobenistant/hr-analytics/principal-component-analysis-explained



import numpy as np  # linear algebra
import pandas as pd  # data rocessing
import matplotlib.pyplot as plt
import seaborn as sns


# 导入数据
# https://www.kaggle.com/nirajvermafcb/d/ludobenistant/hr-analytics/principal-component-analysis-explained
df = pd.read_csv('/home/shen/PycharmProjects/MyPython/Kaggle/human resources/data.csv')



# 了解数据
'''
df.columns.tolist() fetches all the columns and 
then convert it into list type.This step is just 
to check out all the column names in our data.Columns 
are also called as features of our datasets.
'''
columns_names = df.columns.tolist()

print(df.shape)
print(df.head())

# 数值型变量的协方差矩阵
print(df.corr())

# [学习heatmap使用]
# 协方差矩阵的可视化
correlation = df.corr()
plt.figure(figsize=(10,10))
sns.heatmap(correlation, vmax=1, square=True, annot=True, cmap='cubehelix')
plt.title('Correlation between diffrent features')


# 职位
print(df['sales'].unique())

groupby_sales = df.groupby('sales').mean()

IT=groupby_sales['satisfaction_level'].IT
RandD=groupby_sales['satisfaction_level'].RandD
accounting=groupby_sales['satisfaction_level'].accounting
hr=groupby_sales['satisfaction_level'].hr
management=groupby_sales['satisfaction_level'].management
marketing=groupby_sales['satisfaction_level'].marketing
product_mng=groupby_sales['satisfaction_level'].product_mng
sales=groupby_sales['satisfaction_level'].sales
support=groupby_sales['satisfaction_level'].support
technical=groupby_sales['satisfaction_level'].technical



# [学习plt.barh使用]
department_name = df['sales'].unique()
department=(sales, accounting, hr, technical, support, management,
       IT, product_mng, marketing, RandD)

y_pos = np.arange(len(department))
x = np.arange(0, 1, 0.1)

plt.barh(y_pos, department, align='center', alpha=0.8)
plt.yticks(y_pos, department_name)
plt.xlabel('Satisfaction level')
plt.title('Mean Satisfaction Level of each department')



# PCA
# 去除非数值型的变量(feature)
print(df.dtypes)
df_drop = df.drop(labels=['sales', 'salary'], axis=1)
print(df_drop.dtypes)


# [学习如下reindex的步骤]
'''
Here we are converting columns of the dataframe to list 
so it would be easier for us to reshuffle the columns.
We are going to use cols.insert method
'''
cols = df_drop.columns.tolist()
cols.insert(0, cols.pop(cols.index('left')))

df_drop = df_drop.reindex(columns= cols)


# [学习iloc使用]
# 索引分离数据
X = df_drop.iloc[:, 1:8].values
y = df_drop.iloc[:, 0].values

print(np.shape(X))
print(np.shape(y))

# [学习使用sklearn进行标准化]
# 数据的标准化
from sklearn.preprocessing import StandardScaler
X_std = StandardScaler().fit_transform(X)

# [掌握手动计算协方差]
# 计算协方差矩阵
mean_vec = np.mean(X_std, axis=0)
cov_mat = (X_std - mean_vec).T.dot((X_std - mean_vec)) / (X_std.shape[0] - 1)

# another way --> np.cov
print('NumPy covariance matrix: \n%s' %np.cov(X_std.T))


plt.figure(figsize=(8,8))
sns.heatmap(cov_mat, vmax=1, square=True,annot=True,cmap='cubehelix')

plt.title('Correlation between different features')

# [学习使用Numpy简单的线性代数的运算]
# 计算特征值与特征向量
eig_vals, eig_vecs = np.linalg.eig(cov_mat)

print('Eigenvectors \n%s' %eig_vecs)
print('\nEigenvalues \n%s' %eig_vals)



# Make a list of (eigenvalue, eigenvector) tuples
eig_pairs = [(np.abs(eig_vals[i]), eig_vecs[:,i]) for i in range(len(eig_vals))]

# Sort the (eigenvalue, eigenvector) tuples from high to low
eig_pairs.sort(key=lambda x: x[0], reverse=True)

# Visually confirm that the list is correctly sorted by decreasing eigenvalues
print('Eigenvalues in descending order:')
for i in eig_pairs:
    print(i[0])


# Explained Variance
tot = sum(eig_vals)
var_exp = [(i / tot)*100 for i in sorted(eig_vals, reverse=True)]

with plt.style.context('ggplot'):
    plt.figure(figsize=(6, 4))

    plt.bar(range(7), var_exp, alpha=0.5, align='center',
            label='individual explained variance')
    plt.ylabel('Explained variance ratio')
    plt.xlabel('Principal components')
    plt.legend(loc='best')
    plt.tight_layout()


# Projection matrix
# [假设前两个主成分占Explain Variance 的90%]选取前两个变量，构成Projection matrix
matrix_w = np.hstack((eig_pairs[0][1].reshape(7,1), 
                      eig_pairs[1][1].reshape(7,1)
                    ))
print('Matrix W:\n', matrix_w)

# Projection Onto the New Feature Space
Y = X_std.dot(matrix_w)

# [学习使用sklearn做PCA]
# PCA in scikit-learn
from sklearn.decomposition import PCA
pca = PCA().fit(X_std)
plt.plot(np.cumsum(pca.explained_variance_ratio_))
plt.xlim(0, 7, 1)
plt.xlabel('Number of components')
plt.ylabel('Cumulativa explained variance')


from sklearn.decomposition import PCA 
sklearn_pca = PCA(n_components=6)
Y_sklearn = sklearn_pca.fit_transform(X_std)

print(Y_sklearn.shape)

Supplement

补充下sns.heatmap的使用

heatmap(data, vmin=None, vmax=None, cmap=None, center=None, robust=False, annot=None, fmt=’.2g’, annot_kws=None, linewidths=0, linecolor=’white’, cbar=True, cbar_kws=None, cbar_ax=None, square=False, ax=None, xticklabels=True, yticklabels=True, mask=None, **kwargs)

Plot rectangular data as a color-encoded matrix.

This function tries to infer a good colormap to use from the data, but
this is not guaranteed to work, so take care to make sure the kind of
colormap (sequential or diverging) and its limits are appropriate.

This is an Axes-level function and will draw the heatmap into the
currently-active Axes if none is provided to the ax argument. Part of
this Axes space will be taken and used to plot a colormap, unless cbar
is False or a separate Axes is provided to cbar_ax.

import seaborn as sns
import matplotlib.pyplot as plt
flights = sns.load_dataset("flights")
flights = flights.pivot("month", "year", "passengers")
ax = sns.heatmap(flights, linewidth=.5)
plt.show()

输出：

import seaborn as sns
import matplotlib.pyplot as plt
flights = sns.load_dataset("flights")
flights = flights.pivot("month", "year", "passengers")
ax = sns.heatmap(flights, center=flights.loc["January", 1955])
plt.show()

输出：