# 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'

# Import NumPy
import numpy as np

x = np.array([1,2,3,4])
y = np.array([[1,2], [3,4]])
x

y

type(x)

x.shape

y.shape

np.arange(0, 21, 2)

# Try it: Create a numpy array from 0 to 20 in steps of size 2

# Try it: Create a numpy array from -10 to 10 in steps of 0.5 (INCLUDING the number 10)

# Try it: Create a numpy array from 100 to 1000 of length 10

from numpy import random

np.random.randint(1, 100, 50)

# Try it: Create a numpy array filled with random samples 
# from a normal distribution of size 4 x 4

z = np.random.rand(4,4)
z 

z.shape

z.reshape((8,2)) # dim is now 8 x 2

z.flatten()

y = np.array([[1,2], [3,4]])
y.shape

y[0,0] = 10
y 

 # random samples from a uniform distribution between 0 and 1
dat = np.random.rand(4,4)
dat

dat[0, :] # row 1

dat[:, 0] # column 1

dat[0:3:2, 0] # first and third elements in column 1

np.diag(dat) # diagonal

np.arange(32).reshape((8, 4)) # returns an 8 x 4 array

x[0] # returns the first row

def f(x):
    if x >=0:
        return True
    else:
        return False

print f(3)

f_vec = np.vectorize(f)
z = np.arange(-5, 6)
z 

f_vec(z)

def f(x):
    return (x >=0)

print f(3)

# Import SciPy
import scipy

from scipy import stats

from scipy.stats import norm
x = norm.rvs(loc = 0, scale = 1, size = 1000)
plt.hist(x)
plt.title('Histogram of 1000 normal random variables')

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

# DataFrame with 32 observations on 11 variables
mtcars.shape 

# return the column names
mtcars.columns

# return the actual data inside the panadas data frame
mtcars.values

mtcars[25:] # rows 25 to end of data frame

# return index
mtcars.index

mtcars.ix['Maserati Bora'] # access a row by an index

# What other methods are available when working with pandas DataFrames?
# type 'mtcars.' and then click <TAB>
# mtcars.<TAB>

# try it here

mtcars.describe()

(mtcars.mpg >= 20).any()

(mtcars > 0).all()

mtcars['mpg'].hist()
plt.title('Distribution of MPG')
plt.xlabel('Miles Per Gallon')

# Relationship between cyl and mpg
plt.plot(mtcars.cyl, mtcars.mpg, 'o')
plt.xlim(3, 9)
plt.xlabel('Cylinders')
plt.ylabel('MPG')
plt.title('Relationship between cylinders and MPG')

# Relationship between horsepower and mpg
plt.plot(mtcars.hp, mtcars.mpg, 'o')
plt.xlabel('Horsepower')
plt.ylabel('MPG')
plt.title('Relationship between horsepower and MPG')

from pandas.tools.plotting import scatter_matrix
scatter_matrix(mtcars[['mpg', 'hp', 'cyl']], 
               figsize = (10, 6), alpha = 1, diagonal='kde')