data_analysis.py

# -*- coding: utf-8 -*-
"""
Created on Wed Feb 15 00:00:24 2017

@author: Lily
"""

import pandas as pd
import matplotlib.pyplot as plt
import statsmodels.api as sm
import seaborn as sns


df = pd.read_csv('book_data_2.csv')
#clearfy the data
df1 = None
describe = None

def DataCleaning():
    global df1
    df1 = df.dropna(axis=0,how='any')
    
    #delete:0 instead of nan, useless
    print df1[df1['WEB_PRICE'].isin([0.0])]
    df1 = df1.drop([207,402,673,725])
    return df1

#describe the data 
def DescribeData():
    d_p =  df1['PRICE'].describe()
    d_wp = df1['WEB_PRICE'].describe()
    d_s = df1['SCORE'].describe()
    d_pl = df1['PL'].describe()


    d_wp = df1['WEB_PRICE'].describe()


    global describe
    describe = pd.DataFrame({'PRICE':d_p,
                             'PL':d_pl,
                             'WEB_PRICE':d_wp,
                             'SCORE':d_s})
    describe.to_csv('books_describe.csv')

#box plots
def BoxPlots():
    plt.figure()
    df1['SCORE'].plot.box(legend='Score')
    plt.figure()
    df1['PL'].plot.box(legend='PL')
    plt.figure()
    df1['WEB_PRICE'].plot.box(legend='Web_price')
    plt.figure()
    df1['PRICE'].plot.box(legend='Price')


#Histograms
def Histograms():
    plt.figure()
    ax1 = df1['SCORE'].hist(alpha=0.8,bins=50,figsize=(6,4))
    ax1.legend(['Score'])
    plt.figure()
    df1['SCORE'].diff().hist(alpha=0.8,bins=50,figsize=(6,4))


    plt.figure()
    ax2 = df1['PRICE'].hist(alpha=0.8,bins=50,figsize=(6,4))
    ax2.legend(['Price'])
    
    plt.figure()
    df1['PRICE'].diff().hist(alpha=0.8,bins=50,figsize=(6,4))

    plt.figure()
    ax3 = df1['WEB_PRICE'].hist(alpha=0.8,bins=50,figsize=(6,4))
    ax3.legend(['Web_price'])
    plt.figure()
    df1['WEB_PRICE'].diff().hist(alpha=0.8,bins=50,figsize=(6,4))

    plt.figure()
    ax4 = df1['PL'].hist(alpha=0.8,bins=50,figsize=(6,4))
    ax4.legend(['Pl'])
    plt.figure()
    df1['PL'].diff().hist(alpha=0.8,bins=50,figsize=(6,4))


#regression analysis
def RegressionAnalysis():
    
    print df1.corr()
    
    
    sns.lmplot('SCORE','PRICE', df1)
    sns.lmplot('SCORE','WEB_PRICE',df1)
    sns.lmplot('PL','SCORE',df1)


    result1= sm.OLS(df1['SCORE'], df1['PL']).fit()
    print "The result of Score_Pl regression analysis is:"
    print result1.summary()

    result2 = sm.OLS(df1['PRICE'], df1['SCORE']).fit()
    print "The result of Price-Score regression analysis is:"
    print result2.summary()

    result3 = sm.OLS(df1['WEB_PRICE'], df1['SCORE']).fit()
    print "The result of Web_price-Score regression analysis is:"
    print result3.summary()


#Run
DataCleaning()
DescribeData()
BoxPlots()
Histograms()
RegressionAnalysis()