import pandas as pandas
import numpy as np
import scikits.statsmodels.tools.tools as tools
import matplotlib.pyplot as plt
xls = pandas.ExcelFile('/home/pybokeh/Desktop/Civic_Blower.xls')
df = xls.parse('Claims')
ELP_2008 = df[(df['MODEL_YEAR'] == 2008) & (df['FACTORY_CODE'] == 'ELP') &
(df['MODEL_NAME'] == 'CIVIC')]['DAYS_TO_FAIL_MINZERO'].values
HCM_2008 = df[(df['MODEL_YEAR'] == 2008) & (df['FACTORY_CODE'] == 'HCM') &
(df['MODEL_NAME'] == 'CIVIC')]['DAYS_TO_FAIL_MINZERO'].values
SSS_2008 = df[(df['MODEL_YEAR'] == 2008) & (df['FACTORY_CODE'] == 'SSS') &
(df['MODEL_NAME'] == 'CIVIC')]['DAYS_TO_FAIL_MINZERO'].values
ecdf1 = tools.ECDF(ELP_2008)
x1 = np.linspace(min(ELP_2008), max(ELP_2008))
y1 = ecdf1(x1)
ecdf2 = tools.ECDF(HCM_2008)
x2 = np.linspace(min(HCM_2008), max(HCM_2008))
y2 = ecdf2(x2)
ecdf3 = tools.ECDF(SSS_2008)
x3 = np.linspace(min(SSS_2008), max(SSS_2008))
y3 = ecdf3(x3)
fig1 = plt.figure(1)
elp, hcm, sss = plt.step(x1, y1, 'r-', x2, y2, 'b-', x3, y3, 'g-')
plt.grid(True)
plt.ylabel('ECDF', fontsize=8)
df[(df['MODEL_YEAR'] == 2008)
& (df['FACTORY_CODE'] == 'SSS')]['MILES_TO_FAIL'].values,
15,
normed=False,
cumulative=False,
facecolor='green',
alpha=1)
plt.grid(True)
plt.ylabel('Frequency', fontsize=8)
plt.xlabel('MTF Bins', fontsize=8)
plt.title('2008 SSS Civic', fontsize=12)
plt.subplot(224)
dataELP = df[(df['MODEL_YEAR'] == 2008)
& (df['FACTORY_CODE'] == 'ELP')]['MILES_TO_FAIL'].values
ecdf1 = tools.ECDF(dataELP)
x1 = np.linspace(min(dataELP), max(dataELP))
y1 = ecdf1(x1)
dataHCM = df[(df['MODEL_YEAR'] == 2008)
& (df['FACTORY_CODE'] == 'HCM')]['MILES_TO_FAIL'].values
ecdf2 = tools.ECDF(dataHCM)
x2 = np.linspace(min(dataHCM), max(dataHCM))
y2 = ecdf2(x2)
dataSSS = df[(df['MODEL_YEAR'] == 2008)
& (df['FACTORY_CODE'] == 'SSS')]['MILES_TO_FAIL'].values
ecdf3 = tools.ECDF(dataSSS)
x3 = np.linspace(min(dataSSS), max(dataSSS))
y3 = ecdf3(x3)
plt.subplot(2, 1, 2)
n, bins, patches = plt.hist(ELP_CRV_2009,
20,
normed=False,
cumulative=False,
facecolor='red',
alpha=1)
plt.xlabel("MTF")
plt.ylabel("Frequency")
plt.title("ELP CRV MTF Histogram")
plt.grid(True)
# Make a ECDF figure
fig3 = plt.figure(3)
ecdf = tools.ECDF(ELP_CRV_2009)
x = np.linspace(min(ELP_CRV_2009), max(ELP_CRV_2009))
y = ecdf(x)
plt.step(x, y, 'r-')
plt.grid(True)
# Make text figure. Reference: http://matplotlib.sourceforge.net/users/text_props.html
fig4 = plt.figure(4)
ax = fig4.add_axes([0, 0, 1, 1]) # This sets up axis dimensions of 1 x 1
ax.text(0.05,
0.65,
str('2009M ELP CRV:\n' + str(dfCRV_09.describe())),
color='black',
transform=ax.transAxes)
ax.set_axis_off()
plt.xticks(rotation=90)
major_tick_format = FormatStrFormatter('%d')
ax.xaxis.set_major_formatter(major_tick_format)
xtick_labels = ax.xaxis.get_ticklabels()
for label in xtick_labels:
label.set_fontsize(8)
plt.xlabel("DTF")
plt.ylabel("Frequency")
plt.title("2009M TL Keyless Remote Battery DTF Histogram\n"+"Number of bins: " + str(len(bins)-1))
plt.grid(True)
# Make a ECDF figure
fig3 = plt.figure(3)
ecdf2008 = tools.ECDF(DTF2008)
x2008 = np.linspace(min(DTF2008), max(DTF2008))
y2008 = ecdf2008(x2008)
ecdf2009 = tools.ECDF(DTF2009)
x2009 = np.linspace(min(DTF2009), max(DTF2009))
y2009 = ecdf2009(x2009)
plt.step(x2008,y2008,'r-', x2009, y2009, 'b-')
plt.grid(True)
plt.legend(['2008','2009'], loc='best')
plt.title('Empirical CDF vs DTF')
plt.ylabel('% of Failures')
plt.xlabel('Days To Fail')