The writer of this code wants to count the mean and median article length for recent articles on gay marriage in the New York Times. This code has several issues, including errors. When they checked their custom functions against the numpy functions, they noticed some discrepancies. Fix the code so it executes properly, retrieves the articles, and outputs the correct result from the custom functions, compared to the numpy functions.

In [1]:

    

import requests # a better package than urllib2


    

In [23]:

    

def my_mean(input_list):
    list_sum = 0
    list_count = 0
    for el in input_list:
        list_sum += el
        list_count += 1
    return list_sum / list_count

#my_mean([1,2,3]) 
#This code works


    

    
Out[23]:





2.0

In [34]:

    

import math
def my_median(input_list):
    list_length = len(input_list)
    to_use_as_index = math.floor(list_length/2)
    #print(to_use_as_index) #Just checking
    return input_list[to_use_as_index]

#my_median([1,2,3,4,5])


    

In [35]:

    

api_key = "ffaf60d7d82258e112dd4fb2b5e4e2d6:3:72421680"


    

In [36]:

    

url = "http://api.nytimes.com/svc/search/v2/articlesearch.json?q=gay+marriage&api-key=%s" % api_key


    

In [37]:

    

r = requests.get(url)


    

In [60]:

    

wc_list = []
for article in r.json()['response']['docs']:
    if article['word_count'] is not None:
        wc_list.append(article['word_count'])
print(wc_list)


    

    




['25', '920', '576', '868', '1101', '588', '684', '367', '1358']

In [61]:

    

wc_list = [int(i) for i in wc_list]
wc_list.sort(key=int)


    

In [62]:

    

my_mean(wc_list)


    

    
Out[62]:





720.7777777777778

In [63]:

    

import numpy as np


    

In [64]:

    

np.mean(wc_list)


    

    
Out[64]:





720.77777777777783

In [65]:

    

my_median(wc_list)


    

    
Out[65]:





684

In [66]:

    

np.median(wc_list)


    

    
Out[66]:





684.0