In [1]:

    

import numpy as np
from scipy import stats


    

In [2]:

    

import urllib.request as request
file = request.urlopen('https://raw.githubusercontent.com/lidimayra/basic-stats/master/frequencies/pokemons.txt')
print(file)


    

    




<http.client.HTTPResponse object at 0x7fc272070358>

In [3]:

    

pokemons = file.read().decode('UTF-8')


    

In [4]:

    

file.close()


    

In [5]:

    

print(pokemons)


    

    




Pidgeotto
Pidgey
Pidgey
Pidgey
Pidgey
Pidgey
Poliwag
Rapidash
Rattata
Rattata
Sandshrew
Sandshrew

In [6]:

    

pokemons_list = pokemons.split()


    

In [7]:

    

pokemons_list[6]


    

    
Out[7]:





'Poliwag'

In [8]:

    

pokemons_list[7]


    

    
Out[8]:





'Rapidash'

In [9]:

    

pokemons_list[8]


    

    
Out[9]:





'Rattata'

In [10]:

    

stats.mode(pokemons_list)[1]


    

    




/home/lmayra/.pyenv/versions/3.4.3/lib/python3.4/site-packages/scipy/stats/stats.py:257: RuntimeWarning: The input array could not be properly checked for nan values. nan values will be ignored.
  "values. nan values will be ignored.", RuntimeWarning)






    
Out[10]:





array([5])

In [11]:

    

frequencies = stats.itemfreq(pokemons_list)
print(frequencies)


    

    




[['Pidgeotto' '1']
 ['Pidgey' '5']
 ['Poliwag' '1']
 ['Rapidash' '1']
 ['Rattata' '2']
 ['Sandshrew' '2']]

In [12]:

    

type(frequencies)


    

    
Out[12]:





numpy.ndarray

In [35]:

    

xi = frequencies[:, 0]
print(xi)


    

    




['Pidgeotto' 'Pidgey' 'Poliwag' 'Rapidash' 'Rattata' 'Sandshrew']

In [36]:

    

fi = frequencies[:, 1]
print(fi)


    

    




['1' '5' '1' '1' '2' '2']

In [37]:

    

fi = fi.astype(int)
print(fi)


    

    




[1 5 1 1 2 2]

In [38]:

    

np.arange(10)


    

    
Out[38]:





array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9])

In [39]:

    

len(xi)


    

    
Out[39]:





6

In [82]:

    

%matplotlib notebook
import matplotlib.pyplot as plt

size = len(xi)
x_location = np.arange(size)
plt.bar(x_location, fi, align='center')


    

    















    












    
Out[82]:





<Container object of 6 artists>

In [29]:

    

size = len(xi)
print(size)


    

    




6

In [33]:

    

x_location = np.arange(size)
print(x_location)


    

    




[0 1 2 3 4 5]

In [83]:

    

plt.xticks(np.arange(size), xi)


    

    
Out[83]:





([<matplotlib.axis.XTick at 0x7fc26cffd400>,
  <matplotlib.axis.XTick at 0x7fc26d056f28>,
  <matplotlib.axis.XTick at 0x7fc26cf4b7f0>,
  <matplotlib.axis.XTick at 0x7fc26d1752b0>,
  <matplotlib.axis.XTick at 0x7fc26d206b00>,
  <matplotlib.axis.XTick at 0x7fc26d1f4b00>],
 <a list of 6 Text xticklabel objects>)

In [84]:

    

plt.ylim(0, max(fi) + 0.5)


    

    
Out[84]:





(0, 5.5)

In [ ]: