from IPython.core.display import Image
Image(filename='images/ipython-qtconsole.png') 

from IPython.core.display import Image
Image(filename='images/ipython-notebook.png') 

from IPython.core.display import Image
Image(filename='images/spyder-hacking.png') 

%cd <some working directory with the data you want to analyze>

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt

sched_df = pd.read_csv('data/SchedDaysAdv.csv')

sched_df

sched_df.head() # Check out the first few rows of sched_df

sched_df.tail() # Check out the last few rows.

sched_df['ScheduledDaysInAdvance'].describe()

p05_leadtime = sched_df['ScheduledDaysInAdvance'].quantile(0.05)

p05_leadtime

p95_leadtime = sched_df['ScheduledDaysInAdvance'].quantile(0.95)

p95_leadtime

plt.hist(sched_df['ScheduledDaysInAdvance'], 50, normed=1, facecolor='green', alpha=0.75) # normed=1 plots probs instead of counts, alpha in [0,1] is transparency level (RGBA colors)
plt.xlabel('Days')
plt.ylabel('Probability')
plt.title(r'Histogram of Schedule Lead Time')
plt.axis([0, 200, 0, 0.06])
plt.grid(True)
plt.show()

fig1 = plt.figure()
ax1 = fig1.add_subplot(1,1,1)
n, bins, patches = plt.hist(sched_df['ScheduledDaysInAdvance'], 50, normed=1, facecolor='grey', alpha=0.75)

ax1.patch.set_facecolor('#F0F0F0')
ax1.set_title('Histogram of Scheduled Lead Time')
ax1.set_xlabel('Days')
ax1.set_ylabel('Probability')
ax1.grid(True, color='k')
[axp.set_facecolor('white') for axp in ax1.patches] # Seems like there should be a simpler way. Of course, it's easy to just rerun the plt.hist() with the desired color.
                                                   # However, this fine level of control makes it possible to set individual
                                                   # bar colors based on some condition. In fact, when you originally create the histogram and are specifying the
                                                   # color property you can actually set color=[<some list>] where the list contains the colors of each bar. In the
                                                   # process of generating the color list you could do all kinds of logical tests to pick the color of each bar.
display(fig1)

sched_df['ScheduledDaysInAdvance'].hist(bins=50, color='k', alpha=0.3, normed=True)
sched_df['ScheduledDaysInAdvance'].plot(kind='kde', style='k--', xlim=[0,100], title='Histo of Sched Lead Time (using pandas)')


bp = sched_df.boxplot(column='ScheduledDaysInAdvance', by='InsuranceStatus')
fig2 = gcf() # 'g'et 'c'urrent 'f'igure so we can use it later
ax2 = gca()  # 'g'et 'c'urrent 'a'xes so we can use it later

labels = ax2.get_xticklabels()
for label in labels:
    label.set_rotation(90)
display(fig2)

bp = sched_df.boxplot(column='ScheduledDaysInAdvance', by='InsuranceStatus', vert=False)

sched_df_grp1 = sched_df.groupby(['Urgency'])

sched_df_grp1['ScheduledDaysInAdvance'].mean()

sched_df_grp1['ScheduledDaysInAdvance'].quantile(0.95)

sched_df_grp2 = sched_df.groupby(['Urgency','InsuranceStatus'])
sched_df_grp2['ScheduledDaysInAdvance'].mean()

sched_df_grp2['ScheduledDaysInAdvance'].quantile(0.95)

sched_df['ScheduledDaysInAdvance'].hist(bins=50, color='k', alpha=0.3, normed=True, by=sched_df['Urgency'])