# special IPython command to prepare the notebook for matplotlib
%matplotlib inline 

import pandas as pd
import matplotlib.pyplot as plt

pd.options.display.mpl_style = 'default'

url = 'https://raw.githubusercontent.com/cs109/2014_data/master/diamonds.csv'
diamonds = pd.read_csv(url, sep = ',', index_col=0)
diamonds.head()

diamonds.describe()

diamonds['price'].hist(bins=50, color = 'black')
plt.title('Distribution of Price')
plt.xlabel('Price')

# Try changing bin size from 20 to 500. What do you notice? 
diamonds['carat'].hist(bins=20, color = 'black', figsize=(6, 4))
plt.title('Distribution of weights in carats')
plt.xlim(0, 3)
plt.xlabel('Weight in Carats')

diamonds['price'].plot(kind='kde', color = 'black')
plt.title('Distribution of Price')

diamonds.plot(x='carat', y='price', kind = 'scatter', color = 'black', alpha = 1)

plt.scatter(diamonds['carat'], diamonds['price'], color = 'black', alpha = 0.05)
plt.xlabel('Carat')
plt.ylabel('Price')

diamonds.groupby('color').plot(x='carat', y='price', kind = 'scatter', color = 'black', alpha = 1)

# try here

diamonds.boxplot('price', by = 'color')

diamonds.mean()

diamonds.corr() # correlation

diamonds.var() # variance

diamonds.sort('price', ascending = True, inplace = False).head() # sorting

subtable = diamonds.iloc[0:2, 0:2]
print "subtable"
print subtable
print ""

column = diamonds['color']
print "head of the color column"
print column.head()
print ""

row = diamonds.ix[1:2] #row 1 and 2
print "row"
print row
print ""

rows = diamonds.ix[:3] # all the rows before 3
print "rows"
print rows
print ""

color = diamonds.ix[1,'color']
print "color of diamond in row 1"
print color
print ""

# max along column
print "max price %g" % diamonds['price'].max() 
print ""

# axes
print "axes"
print diamonds.axes
print ""

row = diamonds.ix[1]
print "row info"
print row.name
print row.index
print ""

def squared(x):
    """ Return the square of a  
        value """
    return x ** 2

squared(4)

f = lambda x: x**2
f(4)

for i in range(4):
    print 'Hello world!'

prefixes = 'JKLMNOPQ'
suffix = 'ack'
for letter in prefixes:
    print letter + suffix

def countdown(n):
    while n > 0:
        print n
        n = n-1
    print 'Blastoff!'

countdown(3)

a = [5, 10, 15, 20]
b = [i**2 for i in a]
c = [i**2 for i in a if i != 10]

print "a: ", a
print "b: ", b
print "c: ", c