Hy all, I need a little wisdom.
I maage to make a scrapper using the Twitter API and Tweepy. It scrapes tweets from individual profiles. I have a list of around 100 profiles that I want to scrape tweets from, but I cant figure out how to instruct the scraper to extract data from multiple profiles and how to save the output properly in csv. I have the following code:
import tweepy
import time
import pandas as pd
import csv
# API keyws that yous saved earlier
api_key = ''
api_secrets = ''
access_token = ''
access_secret = ''
# Authenticate to Twitter
auth = tweepy.OAuthHandler(api_key,api_secrets)
auth.set_access_token(access_token,access_secret)
#Instantiate the tweepy API
api = tweepy.API(auth, wait_on_rate_limit=True)
username = "markrutte"
no_of_tweets = 3200
try:
#The number of tweets we want to retrieved from the user
tweets = api.user_timeline(screen_name=username, count=no_of_tweets)
#Pulling Some attributes from the tweet
attributes_container = [[tweet.created_at, tweet.favorite_count,tweet.source, tweet.text] for tweet in tweets]
#Creation of column list to rename the columns in the dataframe
columns = ["Date Created", "Number of Likes", "Source of Tweet", "Tweet"]
tweets_df = pd.DataFrame(attributes_container, columns=columns)
except BaseException as e:
print('Status Failed On,',str(e))
time.sleep(3)
In my head, I believe I should specify a list with usernames as the values. And then, for username in list: scrape tweets. However, I dont really know how to do this and am still learning. Can anyone give me some advice or know a tutorial on how I should do this?
Appreciate it.
In my head, I believe I should specify a list with usernames as the values. And then, for username in list: scrape tweets. However, I dont really know how to do this and am still learning. Can anyone give me some advice or know a tutorial on how I should do this?
Appreciate it.
If you put your scraping code into a function, you can then concat its results into an overall dataframe in a loop:
def get_tweets(username, no_of_tweets):
#Creation of column list to rename the columns in the dataframe
columns = ["Date Created", "Number of Likes", "Source of Tweet", "Tweet"]
try:
#The number of tweets we want to retrieved from the user
tweets = api.user_timeline(screen_name=username, count=no_of_tweets)
#Pulling Some attributes from the tweet
attributes_container = [[tweet.created_at, tweet.favorite_count,tweet.source, tweet.text] for tweet in tweets]
# return a dataframe
return pd.DataFrame(attributes_container, columns=columns)
except BaseException as e:
print('Status Failed On,',str(e))
# return an empty dataframe
return pd.DataFrame(columns=columns)
usernames = ['user1', 'user2', 'user3']
no_of_tweets = 3200
tweets_df = pd.concat([get_tweets(username, no_of_tweets) for username in usernames])
Related
Is it possible to exclusively gather Tweets which mention countries by name? I am only gathering Tweets from the US.
I know that Twitter allows us to access context_annotations from the payload, and that context_annotations identifies if a tweet mentions a country. Here, https://developer.twitter.com/en/docs/twitter-api/annotations/overview ,they mention that countries is domain number 160 in context annotations.
I'm wondering if it is possible to exclusively gather Tweets that mention country names. I am not familiar with Tweepy, so I've finally managed to obtain Tweets from the US, but am still unable to specify the code to obtain only tweets which mention countries.
This is my current code:
client = tweepy.Client(bearer_token=bearer_token)
# Specify Query
query = ' "favorite country" place_country:US'
start_time = '2022-03-05T00:00:00Z'
end_time = '2022-03-11T00:00:00Z'
tweets = client.search_all_tweets(query=query, tweet_fields=['context_annotations', 'created_at', 'geo'],
place_fields = ['place_type','geo'], expansions='geo.place_id',
start_time=start_time,
end_time=end_time, max_results=10000)
# Prepare to write to csv file
f = open('tweetSheet.csv','w')
writer = csv.writer(f)
# Write to csv file
for tweet in tweets.data:
print(tweet.text)
print(tweet.created_at)
writer.writerow(['0', tweet.id, tweet.created_at, tweet.text])
# Close csv file
f.close()
has:geo:
One way of doing this would be by filtering in tweets that have country attributes.
You can use the has:geo: operator in your query instead of the place_country: operator seen in the Twitter Docs. This way you get all the tweets that are geo tagged, every geo tagged tweet has a country attribute.
includes
Another way would be checking if the tweet has an includes attribute, empty if it has no geo attributes: response.includes != {}. To get the country code if needed then response.includes['places'][0].country works just fine. It is not very well documented in the Tweepy Docs so here are all the geo attributes found in the Twitter Docs for a tweet:
twt_geo = 1602695447298162689
twt_no_geo = 1602719044645408768
response = client.get_tweet(
twt_geo, place_fields=['country', 'country_code', 'place_type', 'name'], expansions=['geo.place_id'])
if(response.includes != {}):
print(response.includes)
print(response.includes['places'][0].country)
print(response.includes['places'][0].country_code)
print(response.includes['places'][0].place_type)
print(response.includes['places'][0].name)
print(response.includes['places'][0].full_name)
print(response.includes['places'][0])
print(response.data.geo)
print(response.data.geo['place_id'])
else:
print(response.data.id)
Hashtags
If you are implying filtering in tweets that have country names as hashtags as country mentions, you can extract the tweet text with response.data.text and compare the country names you would like to filter in.
I'm trying to get some tweets using Twython, but even with tweet_mode:extended the results are still truncated. Any ideas how I can get the full text.
def requestTweets(topic, resultType = "new", amount = 10, language = "en"):
'''Get the n tweets for a topic, either newest (new) or most popular (popular)'''
#Create Query
query = {'q': topic,
'result_type': resultType,
'count': amount,
'lang': language,
'tweet_mode': 'extended',
}
#Get Data
dict_ = {'user': [], 'date': [], 'full_text': [],'favorite_count': []}
for status in python_tweets.search(**query)['statuses']:
dict_['user'].append(status['user']['screen_name'])
dict_['date'].append(status['created_at'])
dict_['full_text'].append(status['full_text'])
dict_['favorite_count'].append(status['favorite_count'])
# Structure data in a pandas DataFrame for easier manipulation
df = pd.DataFrame(dict_)
df.sort_values(by='favorite_count', inplace=True, ascending=False)
return df
tweets = requestTweets("chocolate")
for index, tweet in tweets.iterrows():
print("***********************************")
print(tweet['full_text'])
Results look like this:
I know it's kind of late! but I put the answer to whom may need it.
I'm using twython==3.9.1 and its possible to get the full text of tweets with the below snippet code:
twython: Twython
tweets = twython.get_user_timeline(
screen_name=user_screen_name, # or user id
count=200, # max count is 200
include_retweets=include_retweets, # could be False or True
exclude_replies=exclude_replies, # could be False or True
tweet_mode='extended', # to get tweets full text
)
I couldn't find a way to do it with Twython, so I switched to tweepy in the end, still, if anyone has an answer, that would be great.
I want to get all tweets for a specific topic that tweeted from NYC. I have twitter credential and my Python code to use the Twitter API is the following:
import tweepy
import numpy as np
import pandas as pd
auth = tweepy.OAuthHandler("......", ".......")
auth.set_access_token("........", "........")
api = tweepy.API(auth)
df = pd.DataFrame(columns = ['Tweets', 'Date of Tweet', 'Retweet Count', 'User Location', 'User Registration Date'])
def stream():
i = 0
for tweet in tweepy.Cursor(api.search, q='climatechange', count=100000, lang='en', tweet_mode='extended', since='2020-02-2', until='2020-02-25',geocode='43.17305,-77.62479,100km').items():
print(i, end='\r')
df.loc[i, 'Tweets'] = tweet.full_text
df.loc[i, 'Date of Tweet'] = tweet.created_at
df.loc[i, 'Retweet Count'] = tweet.retweet_count
df.loc[i, 'User Location'] = tweet.user.location
df.loc[i, 'User Registration Date'] = tweet.user.created_at
df.to_csv('GeoTweets1.csv')
i+=1
if i == 10000:
break
else:
pass
stream()
df.info()
Questions:
1- I want to get metadata. I mean big data of tweets as much as I can. we know that every 15 min we can request 180 keywords and by each request, we can get 100 tweets, which means 18000 tweets. How can I iterate the code that it gives me 18K tweets for a specific keyword and automatically repeat it every 15min? For example, to get tweets of NYC about climatechange, I want to continuously run this code for 10hrs, which equals to 40 of 15mins, which means I can get 720K tweets.
2- I also have the issue to get tweets according to location. When I run the above code and request tweets for a keyword such as climatechanges, it gives me 100 tweets, but for Geo query for New York gives me less. e.g. for geocode='43.17305,-77.62479,32km it gives me 22 tweets and for geocode='43.17305,-77.62479,100km it gives me 12 tweets. Why for Geo search it doesn't give me 100 tweets
Thank you for your help
I am trying to scrape tweets from twitter using twython and I want to use enterprise search api for this because I want to define fromDate and toDate parameters.
I couldn't find any way to do it though, and when I try to cursor tweets from this date, It only returns the tweets about 14 days ago from now.
twitter = Twython(consumer_token, access_token=ACCESS_TOKEN)
# Search parameters
def search_query(QUERY_TO_BE_SEARCHED):
"""
QUERY_TO_BE_SEARCHED : text you want to search for
"""
df_dict=[]
results = twitter.cursor(twitter.search, q=QUERY_TO_BE_SEARCHED,fromDate='2019071200',toDate='2019071400',count=100)
for q in results:
retweet_count = q['retweet_count']
favs_count = q['favorite_count']
date_created = q['created_at']
text = q['text']
hashtags = q['entities']['hashtags']
user_name = '#'+str(q['user']['screen_name'])
user_mentions = []
if(len(q['entities']['user_mentions'])!=0):
for n in q['entities']['user_mentions']:
user_mentions.append(n['screen_name']) # Mentioned profile names in the tweet
temp_dict = {'User ID':user_name,'Date':date_created,'Text':text,'Favorites':favs_count,'RTs':retweet_count,
'Hashtags':hashtags,'Mentions':user_mentions}
df_dict.append(temp_dict)
return pd.DataFrame(df_dict)
that is my code, can you help me improve this ?
I want to look up all the friends (meaning the twitter users one is following) of a sample of friends of one twitter account, to see what other friends they have in common. The problem is that I don't know how to handle protected accounts, and I keep running into this error:
tweepy.error.TweepError: Not authorized.
This is the code I have:
...
screen_name = ----
file_name = "followers_data/follower_ids-" + screen_name + ".txt"
with open(file_name) as file:
ids = file.readlines()
num_samples = 30
ids = [x.strip() for x in ids]
friends = [[] for i in range(num_samples)]
for i in range(0, num_samples):
id = random.choice(ids)
for friend in tweepy.Cursor(api.friends_ids, id).items():
print(friend)
friends[i].append(friend)
I have a list of all friends from one account screen_name, from which I load the friend ids. I then want to sample a few of those and look up their friends.
I have also tried something like this:
def limit_handled(cursor, name):
try:
yield cursor.next()
except tweepy.TweepError:
print("Something went wrong... ", name)
pass
for i in range(0, num_samples):
id = random.choice(ids)
items = tweepy.Cursor(api.friends_ids, id).items()
for friend in limit_handled(items, id):
print(friend)
friends[i].append(friend)
But then it seems like only one friend per sample friend is stored before moving on to the next sample. I'm pretty new to Python and Tweepy so if anything looks weird, please let me know.
First of all, a couple of comments on naming. The names file and id are protected, so you should avoid using them to name variables - I have changes these.
Secondly, when you initialise your tweepy API, it's clever enough to deal with rate limits if you use wait_on_rate_limit=True and will inform you when it's delayed due to rate limits if you use wait_on_rate_limit_notify=True.
You also lose some information when you set friends = [[] for i in range(num_samples)], as you then won't be able to associate the friends you find with the account they relate to. You can instead use a dictionary, which will associate each ID used with the friends found, allowing for better processing.
My corrected code is as follows:
import tweepy
import random
consumer_key = '...'
consumer_secret = '...'
access_token = '...'
access_token_secret = '...'
# OAuth process, using the keys and tokens
auth = tweepy.OAuthHandler(consumer_key, consumer_secret)
auth.set_access_token(access_token, access_token_secret)
# Creation of the actual interface, using authentication. Use rate limits.
api = tweepy.API(auth, wait_on_rate_limit=True, wait_on_rate_limit_notify=True)
screen_name = '----'
file_name = "followers_data/follower_ids-" + screen_name + ".txt"
with open(file_name) as f:
ids = [x.strip() for x in f.readlines()]
num_samples = 30
friends = dict()
# Initialise i
i = 0
# We want to check that i is less than our number of samples, but we also need to make
# sure there are IDs left to choose from.
while i <= num_samples and ids:
current_id = random.choice(ids)
# remove the ID we're testing from the list, so we don't pick it again.
ids.remove(current_id)
try:
# try to get friends, and add them to our dictionary value if we can
# use .get() to cope with the first loop.
for page in tweepy.Cursor(api.friends_ids, current_id).pages():
friends[current_id] = friends.get(current_id, []) + page
i += 1
except tweepy.TweepError:
# we get a tweep error when we can't view a user - skip them and move onto the next.
# don't increment i as we want to replace this user with someone else.
print 'Could not view user {}, skipping...'.format(current_id)
The output is a dictionary, friends, with keys of user IDs and items of the friends for each user.