Is Donald Trump Kanye West?

In [ ]:

    

!pip install tweepy
import tweepy

consumer_key = "XXXXXX"
consumer_secret = "YYYYY"

access_token = "ZZZZZZZ"
access_token_secret = "AAAAAAA"


    

In [ ]:

    

userList = ['realdonaldtrump','kanyewest','katyperry','justinbieber','taylorswift13','BarackObama','rihanna','YouTube','ladygaga','TheEllenShow','twitter','jtimberlake','KimKardashian','britneyspears','Cristiano','selenagomez','jimmyfallon','cnnbrk','ArianaGrande','instagram','shakira','ddlovato','JLo','Drake','Oprah','KingJames','KevinHart4real','MileyCyrus','BillGates','onedirection','nytimes','SportsCenter','espn','Harry_Styles','Pink','LilTunechi','CNN','wizkhalifa','Adele','NiallOfficial','BrunoMars','KAKA','ActuallyNPH','kany,west','BBCBreaking','danieltosh','neymarjr','aliciakeys','LiamPayne','Louis_Tomlinson','NBA','EmWatson','pitbull','narendra,odi','SrBachchan','khloekardashian','ConanOBrien','iamsrk','Eminem','kourtneykardash','NICKIMINAJ','realmadrid','davidguett,','AvrilLavigne','NFL','zaynmalik','KendallJenner','BeingSalmanKhan','FCBarcelona','aamir_khan','NASA','blakeshelton','Kyli,Jenner','aplusk','coldplay','vine','chrisbrown','edsheeran','MariahCarey','xtina','LeoDiCaprio','agnezmo','TwitterEspanol','deepikapadukone','MohamadAlarefe','BBCWorld','google','TheEconomist','JimCarrey','shugairi','KDTrey5','ivetesangalo','priyankachopra','RyanSe,crest','iHrithik','Beyonce','TwitterSports','SnoopDogg','Reuters','AlejandroSanz','ricky_martin','radityadika']


    

In [ ]:

    

from collections import namedtuple


def get_all_tweets(screen_name):
	#Twitter only allows access to a users most recent 3240 tweets with this method
	
	#authorize twitter, initialize tweepy
	auth = tweepy.OAuthHandler(consumer_key, consumer_secret)
	auth.set_access_token(access_token, access_token_secret)
	api = tweepy.API(auth)
	
	#initialize a list to hold all the tweepy Tweets
	alltweets = []	
	
	#make initial request for most recent tweets (200 is the maximum allowed count)
	new_tweets = api.user_timeline(screen_name = screen_name,count=200)
	
	#save most recent tweets
	alltweets.extend(new_tweets)
	
	#save the id of the oldest tweet less one
	oldest = alltweets[-1].id - 1
	
	#keep grabbing tweets until there are no tweets left to grab
	while len(new_tweets) > 0:
		#print "getting tweets before %s" % (oldest)
		#all subsiquent requests use the max_id param to prevent duplicates
		new_tweets = api.user_timeline(screen_name = screen_name,count=200,max_id=oldest)		
		#save most recent tweets
		alltweets.extend(new_tweets)		
		#update the id of the oldest tweet less one
		oldest = alltweets[-1].id - 1		
		#print "...%s tweets downloaded so far" % (len(alltweets))	
	#transform the tweepy tweets into a 2D array that will populate the csv	
	outtweets = [[tweet.id_str, tweet.created_at, tweet.text.encode("utf-8")] for tweet in alltweets]
	
	#write the csv	
	with open('%s_tweets.csv' % screen_name, 'wb') as f:
		writer = csv.writer(f)
		writer.writerow(["id","created_at","text"])
		writer.writerows(outtweets)
	pass

for user in userList:
   get_all_tweets(user)


    

In [ ]:

    

#compare the bin heights of two histograms
#this particular implementation uses the intersection as a similarity metric
def compHist(first, second):
    i=0
    r=0
    while i< len(first):
        r = r + min(first[i],second[i])
        i=i+1
    return r

#get tweets per day
def getTPD(df):
    dates = [datetime.datetime(int(x[0:4]),int(x[5:7]),int(x[8:10])) for x in df['created_at'].values] #translates a string to a date object
    datesU = pd.unique(dates)
    cc = np.array([])
    for D in datesU:
        c=0
        for D2 in dates:
            if D2==D:
                c=c+1
        cc = np.append(cc,c)
    return [cc.mean(), cc.std()]
        
#a "narcicism" index.  Basically counting up occurences of "I" and "me"
def getNarcInd(df):
    P = re.compile(r"I[\'m]*[\s]|\sme$|\sme[\s\.]")
    A = np.array([len(P.findall(x)) for x in df['text'].values])
    #df['narcInd'] = A
    return A.mean()


    

In [ ]:

    

tgt = datetime.datetime(2016,8,26)-datetime.timedelta(90)
celebStruct = namedtuple("celebStruct", "user numTweets histA tweetsPerDay narcInd")
rr=[]

for user in userList:
    df = pd.read_csv(user + '_tweets.csv')
    
    #get most recent 3 months
    recentTgt =[datetime.date(int(x[0:4]),int(x[5:7]),int(x[8:10]))>tgt.date() for x in df['created_at']]
    recentTweets = df[recentTgt]

    numTweets = len(recentTweets ) #tweets per 3 months

    #create a histogram for each celebrity indicating tweets per hour
    bin1 = range(25)
    fig = plt.plot()    
    hourInformation = [int(x[11:13]) for x in df['created_at']] #strip off time information, leaving only the hour of day
    [A,B,C] =plt.hist((hourInformation), normed=True,bins = bin1)
    #the next two lines will plot the histogram with consistent axes, and save each one, allowing for visual inspection
    plt.ylim([0,.1])
    plt.savefig(user + '_hist.png')
    plt.close()
    
    #two random measures I made up.  Not sure how important they are!
    narcisistIndex=getNarcInd(recentTweets)    
    tweetsPerDay = getTPD(recentTweets)

    #create a new object, adds it to the list of Twitter accounts
    thisCeleb = celebStruct(user, numTweets, linksTweeted, A,tweetsPerDay,narcisistIndex)
    rr.append(thisCeleb)


    

In [ ]:

    

TRUMP = rr[0]
#this portion of the code computes and compares histograms
#of what hour of day people tweet, against the histogram of Donald Trump's daily tweeting frequency.
i=0
m_FINAL =[]
while i<len(rr):
    current = rr[0]
    m_FINAL.append([rr[i].user ,compHist(TRUMP.histA, rr[i].histA)])
    i=i+1
print m_FINAL
L = pd.DataFrame(m_FINAL,columns=['name','match'])
L.sort('match')


    

In [ ]:

    




    

In [ ]:

    

[plt.plot(x[4],'o',color='#3366aa') for x in rr]


    

In [ ]:

    

A = [[x[0],x[4]] for x in rr]


    

In [ ]: