Exploring the basics of the Twitter API with Twython¶

Instantiate an API client in Twython¶

Again, this section assumes you’ve managed to get through the previous guide:

Creating a Twitter Application for Programmatic Access with Twython

If you’ve followed the app creation/authentication steps, then I’m assuming that you can instantiate a Twitter API client in Twython in an interactive Python shell, like this:

from twython import Twython
import json

CREDENTIALS_FILENAME = 'creds-twitter.json'
jf = open(CREDENTIALS_FILENAME)
creds = json.load(jf)
jf.close()

client = Twython(creds['consumer_key'],
                  creds['consumer_secret'],
                  creds['access_token'],
                  creds['access_token_secret'])

Of course, adjust the above snippet for your own situation, if you don’t happen to have saved your user/app credentials in a text file named creds-twitter.json, the contents of which look like:

{
   "screen_name": 'yoursecondaryaccount',
   "consumer_key": "abcYOURCONSUMERKEY",
   "consumer_secret": "abcYOURCONSUMERSECRET",
   "access_token": "abcYOURACCESSTOKEN",
   "access_token_secret": "abcYOURACCESSTOKENSECRET"
}

Check Twitter with your web browser¶

As with everything that is programmatic, there is an “old-fashioned” way to do things. Open up a web browser and visit Twitter. At the very least, we can confirm that your Internet is working and that Twitter is operating normally.

While we’re here, let’s reflect why web pages exist: to display information in a (hopefully) universally attractive and accessible way. This means that no matter how dense the Twitter data objects get, we can always refer to the view created by Twitter’s talented web designers to know what are the most important parts of the data. Those web designers aren’t cheap.

Let’s focus on an instance of a Twitter data object that, in the year 2017, is well known: the profile page and tweets of @realDonaldTrump.

The profile page is accessible at:

https://twitter.com/realDonaldTrump

It is also accessible at:

https://twitter.com/realdonaldtrump

That both URLs resolve to the same user gives us some insight to how Twitter tracks Twitter names – specifically, that they are case-insensitive. But that is a relatively minor detail compared to how @realdonaldtrump’s information is rendered for your web browser.

Here is a screenshot:

Keep your browser window open. The rest of this guide is an investigation to how Twitter “objectifies” – aka serializes – the real-world force-of-nature known to the world as @realDonaldTrump.

Twython API methods return dictionaries¶

Again, we want to know: how does Twitter [data-]objectify the real-world entity of @realDonaldTrump? Or, to put it another way, in what form does @realDonaldTrump exist as Twitter transmits it to the millions/billions of applications, i.e. web browsers and Twitter phone apps?

Like most modern APIs, the Twitter API uses the JSON format to serialize and transmit data. Twython API methods – such as show_user – contact the Twitter API, receive the data as JSON, and automatically, for our convenience, deserialize it as a Python dictionary.

So, in the previous snippet, the result variable points to a Python dict, which encapsulates all the information Twitter chooses to reveal about @realdonaldtrump.

It’s a pretty big dictionary, because show_user returns not just data about the user, but data about the user’s latest tweet.

"description": "45th President of the United States of America",
"friends_count": 43,
"statuses_count": 34549,
"created_at": "Wed Mar 18 13:46:38 +0000 2009",
"screen_name": "realDonaldTrump",
"listed_count": 65579,
"name": "Donald J. Trump",
"id": 25073877,
"location": "Washington, DC",
"profile_image_url_https": "https://pbs.twimg.com/profile_images/1980294624/DJT_Headshot_V2_normal.jpg",
"favourites_count": 45,
"followers_count": 25812629,
"protected": false,
"verified": true

"id": 25073877,
"notifications": false,
"following": false,

Exploring @realdonaldtrump latest Tweet object¶

The latest tweet by @realdonaldtrump is itself a dictionary object, accessed by the user object’s "status" key.

Here is the direct URL to that status (i.e. tweet):

https://twitter.com/realDonaldTrump/status/836933725602656256

A screenshot:

Here is an excerpt of its most important key/value pairs:

"status": {
  "favorite_count": 156827,
  "favorited": false,
  "text": "THANK YOU!",
  "is_quote_status": false,
  "lang": "en",
  "source": "<a href=\"http://twitter.com/download/android\" rel=\"nofollow\">Twitter for Android</a>",
  "retweeted": false,
  "created_at": "Wed Mar 01 13:38:44 +0000 2017",
  "truncated": false,
  "retweet_count": 24172,
  "in_reply_to_screen_name": null,
  "in_reply_to_status_id": null,
  "id": 836933725602656256,
  "in_reply_to_user_id": null,
  "geo": null
},

Again, some of the fields are confusing without knowing the context of how we’re accessing the API on behalf of the authenticated user. Though this tweet has been retweeted more than 24,000 times, the 'retweeted' value is false, because the user represented by client has not yet retweeted (or favorited) this tweet.

To reiterate the concept of data-as-Python-dict, step through the data object using Python expressions:

>>> result['name']
'Donald J. Trump'
>>> result['screen_name']
'realDonaldTrump'
>>> result['description']
'45th President of the United States of America'
>>> result['created_at']
'Wed Mar 18 13:46:38 +0000 2009'
>>> result['status']['text']
'THANK YOU!'
>>> result['status']['created_at']
'Wed Mar 01 13:38:44 +0000 2017'

And might as well try some calculations. What is the ratio of favorites on that tweet versus retweets?

>>> tweet = result['status']
>>> tweet['favorite_count'] / tweet['retweet_count']
6.48796127751117

How many tweets per day, on average, has @realdonaldtrump tweeted since its inception? The 'created_at' field is a human-readable datetime string – 'Wed Mar 18 13:46:38 +0000 2009' – so we need to convert it to a datetime object before doing the arithmetic:

>>> from datetime import datetime
>>> from dateutil import parser
>>> xtime = parser.parse(result['created_at']).timestamp()
>>> tsecs = datetime.now().timestamp() - xtime
>>> tdays = tsecs / (60 * 60 * 24)
>>> result['statuses_count'] / tdays
11.889841748974751

Twython API methods map to documented Twitter API endpoints¶

The Twython library is meant to be a convenience. Rather than having to know exactly the name of the Twitter API endpoint that gets us user profile information, and then accessing that endpoint (e.g. using the Requests library) we just have to know about Twython’s show_user method, which executes the HTTP requests, receives the data as raw text, and deserializes it for us.

The Twython authors have decided to make their methods follow very closely to the actual names and conventions of Twitter’s documented API endpoints.

In the case of show_user, the corresponding Twitter endpoint is known as users/show.

Its documentation is at:

https://dev.twitter.com/docs/api/1.1/get/users/show

For each Twython method in Twython’s source code, the authors conveniently list the corresponding Twitter API URL.

Twython’s documentation for the show_user method can be viewed in your web browser at this URL:

https://twython.readthedocs.io/en/latest/api.html#twython.Twython.show_user

But if you’re trying Twython out in ipython (which you should be), add a question mark after a method name to quickly see this info:

::emphasize-lines: 1,4

>>> client.show_user?
Signature: client.show_user(**params)
Docstring:
Returns a variety of information about the user specified by the
required user_id or screen_name parameter.

Docs: https://dev.twitter.com/docs/api/1.1/get/users/show
File:      ~/.pyenv/versions/anaconda3-4.1.1/lib/python3.5/site-packages/twython/endpoints.py
Type:      method

When I say that the arguments of a Twython method map to the parameters of API endpoint, I mean that the Twython methods have been designed to use arguments named like the parameters as defined by Twitter.

For the users/show endpoint, look at the Parameters section:

https://dev.twitter.com/rest/reference/get/users/show#parameters

There are 3 documented parameters: screen_name, which we just used, and user_id and show_entities.

What is the difference between a user’s ID and screen name?¶

Like most modern data systems, such as the Social Security Administration or 4chan, data objects in Twitter are tracked internally with unique numbers, independent of their real-world human-readable identifiers (i.e. your “name”).

Each Twitter account is assigned a unique integer that never changes no matter how often a Twitter user changes their screen name, so that an account’s history can be consistently tracked no matter how it changes its human-readable name .

For example, part of the presidential transition from Obama to Trump involved creating new social media accounts for the Trump administration. Before January 20, 2017, the Twitter account known as @whitehouse was renamed to @obamawhitehouse. And a new account, with an empty Twitter history (but an auto-following of the Obama @whitehouse 13+ million followers) was created and given the screen name of @whitehouse.

Additional reading from Recode: Donald Trump is getting a second Twitter account and 13 million new followers Friday

Don’t take my word for it, use Twython and show_user to confirm it for yourself:

>>> w44 = client.show_user(screen_name='obamawhitehouse')
>>> w45 = client.show_user(screen_name='whitehouse')
>>> w44['created_at']
'Fri Apr 10 21:10:30 +0000 2009'
>>> w45['created_at']
'Thu Jan 19 22:54:27 +0000 2017'

Remember when we retrieved the data for the @realdonaldtrump user? That data object contains the user ID for @realdonaldtrump:

>>> u = client.show_user(screen_name='realdonaldtrump')
>>> u['id']
25073877

The easiest test of the user_id parameter for the show_user method is to see if we can get the @realdonaldtrump data object by referring to its ID: 25073877

>>> v = client.show_user(user_id=25073877)
>>> v['screen_name']
'realDonaldTrump'

OK, that was boring. Let’s try an arbitrary user ID:

>>> x = client.show_user(user_id=1000)
>>> x['screen_name']
'curiousjordy'
>>> x['created_at']
'Sat Jul 15 08:41:38 +0000 2006'

Wow, user ID of 1000 was created substantially earlier than @realdonaldtrump, i.e. user 25073877. It seems logical that ID 1 corresponds to Twitter’s first user...

>>> client.show_user(user_id=1)
TwythonError: Twitter API returned a 404 (Not Found), User not found.

A TwythonError – get used to seeing these. We’ll see them when Twitter is down. When we’ve surpassed or API access rate limits. Or when we ask for nonsensical or non-existent values.

I’m going to hand-wave past the very complex concept of “error handling” for now (i.e. not by the end of the typical quarter in “learning how to program for the first time”)

But for those of you familiar with the concept, if we wanted to find the smallest Twitter user ID that belongs to a retrievable Twitter user, here’s one approach using error handling:

from twython import TwythonError
xcount = 0

while True: # don't stop till we get a valid user!
  try:
    xcount += 1
    user = client.show_user(user_id=xcount)
  except TwythonError:
    print("No user for ID of:", xcount)
    continue
  else:
    # successful query
    tmptxt = "User ID of {id} belongs to user named {sname}"
    print(tmptxt.format(id=user['id'], sname=user['screen_name']))
    # must get out of the infinite while loop
    break

The result of that routine:

No user for ID of: 1
No user for ID of: 2
No user for ID of: 3
No user for ID of: 4
No user for ID of: 5
No user for ID of: 6
No user for ID of: 7
No user for ID of: 8
No user for ID of: 9
No user for ID of: 10
No user for ID of: 11
User ID of 12 belongs to user named jack

Who is @jack? The co-founder and a CEO of Twitter:

https://en.wikipedia.org/wiki/Jack_Dorsey

https://twitter.com/jack

A screenshot of his profile:

Tweeted URLs vs. shortened URLs¶

If you look at the screenshot of @jack’s tweet, you’ll notice that the text is rendered as:

Making progress every day. We’ll fix this! https://www.buzzfeed.com/amphtml/charlie …

But what we see is not what we get – clicking that link will take you to the fully-resolved URL:

https://www.buzzfeed.com/amphtml/charliewarzel/twitter-tackles-trolls-with-three-new-anti-harassment-featur

But try right-clicking the link in your browser, and selecting the “Copy link address” action:

Check your clipboard to see the copied value. It is the URL:

https://t.co/wRiFAsjmbN

Long story short, everytime you include a URL in a tweet, Twitter automatically passes it through its URL shortening service. Back in the old days, URL shorteners were used so that really long URLs wouldn’t completely eat up Twitter’s 140-character-per-Tweet limit. That is, users would have to visit a site like https://bit.ly, paste in a long URL, copy the shortened URL, and include use that in their tweet. The URL shortening service – in this case, bit.ly – does the work of translating the short URL to its intended destination.

Twitter’s t.co service was created to cut out the middleman. And also, for Twitter to have more analytics. Everytime someone clicks on a t.co link, Twitter is able to count the total times that short link has been clicked, and information about each individual “click”.

In terms of programmatic access, reading the tweet object’s text value will only get you the shortened version of a URL:

>>> tweet = client.show_status(id=828982363610755073)
>>> tweet['text']
"Making progress every day. We'll fix this! https://t.co/wRiFAsjmbN"

Thus, if you want to do an analysis of what sites a Twitter user tweets – like the analysis done by BuzzFeed on Here’s Where Donald Trump Gets His News – you’ll have to do one of two things:

1. Access the "entities" data of a tweet

The tweet object has several nested dictionaries; "entities" is where metadata for the tweeted URLs is stored:

"entities": {
      "symbols": [],
      "urls": [
        {
          "expanded_url": "https://www.buzzfeed.com/amphtml/charliewarzel/twitter-tackles-trolls-with-three-new-anti-harassment-featur",
          "display_url": "buzzfeed.com/amphtml/charli\u2026",
          "url": "https://t.co/wRiFAsjmbN",
          "indices": [
            43,
            66
          ]
        }
      ],
      "hashtags": [],
      "user_mentions": []
    },

For any given tweet, here’s how to print the expanded URLs (if any) of each URL, as far as Twitter is concerned:

for url in tweet['entities']['urls']:
    print(url['expanded_url'])

2. Expand the URL using good-old fashioned HTTP principles

In Twitter’s older data, however, there aren’t “expanded” tweets. Or, even today, people will tweet a shortened-link, and Twitter will still apply its own t.co shortening. For example, take a look at this Trump tweet:

https://twitter.com/realDonaldTrump/status/829451566000242688

Trump tweeted a pre-shortened link, which Twitter shortened to:

https://t.co/6LmsR5JOSW

The “expanded” version of that Tweet is what Trump originally tweeted:

http://bit.ly/2k4b0imEmersonPoll

>>> xt = client.show_status(id=829451566000242688)
>>> xt['entities']['urls'][0]

{'display_url': 'bit.ly/2k4b0imEmerson…',
 'expanded_url': 'http://bit.ly/2k4b0imEmersonPoll',
 'indices': [61, 84],
 'url': 'https://t.co/6LmsR5JOSW'}

What we really want is what http://bit.ly/2k4b0imEmersonPoll redirects to:

http://thehill.com/homenews/media/318514-trump-admin-seen-as-more-truthful-than-news-media-poll

Handling URL redirects is all part of the plain, HTTP spec. Which means we can use a general web-requesting library, like Requests, to get the final URL:

http://docs.python-requests.org/en/master/user/quickstart/#redirection-and-history

>>> import requests
>>> resp = requests.get('http://bit.ly/2k4b0imEmersonPoll')
>>> resp.url
'http://thehill.com/homenews/media/318514-trump-admin-seen-as-more-truthful-than-news-media-poll'

I highly recommend reading that BuzzFeed story – Here’s Where Donald Trump Gets His News – it’s an interesting example of what you can do when mining Twitter for data is not just an end to itself, but as a proxy into more data sources and a bigger story.

A tweet that is a reply¶

How does Twitter serialize the tweets that are attached as “replies”?

Only one way to find out. Find a tweet that has been replied to, and get that reply-tweet’s ID. Let’s refer to another tweet by @jack at the following URL:

https://twitter.com/jack/status/836959185220710402

Screenshot of that tweet, which includes the rendering of its most direct/prominent reply:

You might have noticed that this tweet by @jack is itself a reply to a previous tweet by @jack. What’s that about? It’s a common convention for Twitter users to reply to their own tweets when they have a stream-of-consciousness/rant that requires more than 140 characters. Because the Twitter web (and phone) client stiches the original tweet and replies into a single page, it’s a good workaround for Twitter’s individual tweet limit.

But let’s look at the first shown reply by a non-@jack user – the tweet-reply by @HilzFuld has this direct URL:

https://twitter.com/HilzFuld/status/836964090245447680

Let’s get @HilzFuld’s reply as a data object in Twython:

replytweet = client.show_status(id=836964090245447680)

I’ve included the full object as JSON in the appendix below:

twitter_hilzfuld_reply_to_jack_tweet

For our purposes, it’s worth noting that the reply-related fields have values instead of null:

>>> replytweet['in_reply_to_screen_name']
'jack'
>>> replytweet['in_reply_to_status_id']
836959185220710402
>>> replytweet['in_reply_to_user_id']
12

Worth noting when trying to figure out how to send a reply tweet via Twython.

Cursors¶

(for advanced users)

Twitter limits the maximum number of tweets you can get from a user timeline in a single request: 200. However, it allows you to go back to the past 3,200 tweets. This requires understanding the concept of a Twitter “cursor”, or pagination:

https://twython.readthedocs.io/en/latest/usage/advanced_usage.html#search-generator

https://twython.readthedocs.io/en/latest/usage/special_functions.html#cursor

Replying to a tweet¶

Tweets that are meant to be a reply to another tweet are serialized and rendered differently than plain tweets, as we saw in a previous section: A tweet that is a reply.

Sending a tweet meant as a reply uses the update_status endpoint, but requires the filling out an additional argument, and adding something to the status argument.

• We must provide the in_reply_to_status_id argument, which is the id number of the tweet to which we’re replying.
• The status argument, which is the content of our reply, must include the screen name of the user to whom we are replying.

Given this tweet: https://twitter.com/realDonaldTrump/status/836933725602656256

Our update_status call looks like this:

client.update_status(status="You are SOOOO welcome @realdonaldtrump!",
                     in_reply_to_status_id=836933725602656256)

What happens when you don’t follow Twitter’s requirements for a “reply”? According to the statuses/update documentation, the in_reply_to_status_id parameter is “ignored unless the author of the Tweet this parameter references is mentioned within the status text.”

Test it out for yourself: you’ll see that the tweet will be posted, but it will not be considered as a reply to any tweet:

client.update_status(status="okey dokey annie oakley",
                 in_reply_to_status_id=836933725602656256)

client.update_status(status="<3 u @realdonaldtrump",
                     in_reply_to_status_id=1)

def reply_to_tweet(client, tweet, reply_text="OK"):
    tweet_id = tweet['id']
    author_name = tweet['user']['screen_name']
    reply = "Hey @{}: {}".format(author_name, reply_text)
    r = client.update_status(status=reply, in_reply_to_status_id=tweet_id)
    return r


sometweet = client.show_status(id=836794907855552518)

reply_to_tweet(client, sometweet, 'Loved that! MOAR plz')

Full JSON result for users/show, @realdonaldtrump¶

{
  "id_str": "25073877",
  "profile_background_image_url": "http://pbs.twimg.com/profile_background_images/530021613/trump_scotland__43_of_70_cc.jpg",
  "profile_text_color": "333333",
  "description": "45th President of the United States of America",
  "utc_offset": -18000,
  "is_translator": false,
  "friends_count": 43,
  "url": null,
  "profile_sidebar_fill_color": "C5CEC0",
  "translator_type": "regular",
  "statuses_count": 34549,
  "created_at": "Wed Mar 18 13:46:38 +0000 2009",
  "has_extended_profile": false,
  "profile_background_image_url_https": "https://pbs.twimg.com/profile_background_images/530021613/trump_scotland__43_of_70_cc.jpg",
  "screen_name": "realDonaldTrump",
  "contributors_enabled": false,
  "status": {
    "id_str": "836933725602656256",
    "place": null,
    "favorite_count": 156827,
    "coordinates": null,
    "entities": {
      "symbols": [],
      "urls": [],
      "hashtags": [],
      "user_mentions": []
    },
    "in_reply_to_user_id_str": null,
    "favorited": false,
    "text": "THANK YOU!",
    "contributors": null,
    "is_quote_status": false,
    "lang": "en",
    "source": "<a href=\"http://twitter.com/download/android\" rel=\"nofollow\">Twitter for Android</a>",
    "retweeted": false,
    "created_at": "Wed Mar 01 13:38:44 +0000 2017",
    "truncated": false,
    "in_reply_to_status_id_str": null,
    "retweet_count": 24172,
    "in_reply_to_screen_name": null,
    "in_reply_to_status_id": null,
    "id": 836933725602656256,
    "in_reply_to_user_id": null,
    "geo": null
  },
  "follow_request_sent": false,
  "profile_background_color": "6D5C18",
  "default_profile": false,
  "profile_location": null,
  "following": false,
  "listed_count": 65579,
  "name": "Donald J. Trump",
  "id": 25073877,
  "entities": {
    "description": {
      "urls": []
    }
  },
  "location": "Washington, DC",
  "profile_image_url_https": "https://pbs.twimg.com/profile_images/1980294624/DJT_Headshot_V2_normal.jpg",
  "is_translation_enabled": true,
  "notifications": false,
  "lang": "en",
  "profile_use_background_image": true,
  "geo_enabled": true,
  "favourites_count": 45,
  "profile_sidebar_border_color": "BDDCAD",
  "followers_count": 25812629,
  "time_zone": "Eastern Time (US & Canada)",
  "profile_image_url": "http://pbs.twimg.com/profile_images/1980294624/DJT_Headshot_V2_normal.jpg",
  "profile_banner_url": "https://pbs.twimg.com/profile_banners/25073877/1485301108",
  "profile_background_tile": true,
  "protected": false,
  "profile_link_color": "0D5B73",
  "default_profile_image": false,
  "verified": true
}

Full JSON result for show/statuses/828982363610755073¶

{
  "id_str": "828982363610755073",
  "place": null,
  "favorite_count": 268,
  "user": {
    "following": true,
    "statuses_count": 21453,
    "id_str": "12",
    "listed_count": 27171,
    "name": "jack",
    "default_profile_image": false,
    "is_translator": false,
    "profile_background_color": "EBEBEB",
    "entities": {
      "description": {
        "urls": []
      }
    },
    "translator_type": "regular",
    "description": "",
    "utc_offset": -28800,
    "profile_image_url_https": "https://pbs.twimg.com/profile_images/768529565966667776/WScYY_cq_normal.jpg",
    "protected": false,
    "is_translation_enabled": false,
    "friends_count": 2538,
    "notifications": false,
    "lang": "en",
    "url": null,
    "profile_sidebar_fill_color": "F3F3F3",
    "profile_use_background_image": true,
    "geo_enabled": true,
    "favourites_count": 16257,
    "location": "California, USA",
    "created_at": "Tue Mar 21 20:50:14 +0000 2006",
    "profile_sidebar_border_color": "DFDFDF",
    "has_extended_profile": true,
    "followers_count": 3986061,
    "profile_background_image_url_https": "https://abs.twimg.com/images/themes/theme7/bg.gif",
    "screen_name": "jack",
    "time_zone": "Pacific Time (US & Canada)",
    "contributors_enabled": false,
    "profile_image_url": "http://pbs.twimg.com/profile_images/768529565966667776/WScYY_cq_normal.jpg",
    "profile_banner_url": "https://pbs.twimg.com/profile_banners/12/1483046077",
    "profile_background_tile": false,
    "follow_request_sent": false,
    "id": 12,
    "profile_link_color": "990000",
    "profile_background_image_url": "http://abs.twimg.com/images/themes/theme7/bg.gif",
    "profile_text_color": "333333",
    "default_profile": false,
    "verified": true
  },
  "coordinates": null,
  "entities": {
    "symbols": [],
    "urls": [
      {
        "expanded_url": "https://www.buzzfeed.com/amphtml/charliewarzel/twitter-tackles-trolls-with-three-new-anti-harassment-featur",
        "display_url": "buzzfeed.com/amphtml/charli\u2026",
        "url": "https://t.co/wRiFAsjmbN",
        "indices": [
          43,
          66
        ]
      }
    ],
    "hashtags": [],
    "user_mentions": []
  },
  "in_reply_to_user_id_str": null,
  "favorited": false,
  "text": "Making progress every day. We'll fix this! https://t.co/wRiFAsjmbN",
  "contributors": null,
  "is_quote_status": false,
  "lang": "en",
  "source": "<a href=\"http://twitter.com/download/iphone\" rel=\"nofollow\">Twitter for iPhone</a>",
  "retweeted": false,
  "possibly_sensitive": false,
  "created_at": "Tue Feb 07 15:02:51 +0000 2017",
  "truncated": false,
  "in_reply_to_status_id_str": null,
  "retweet_count": 61,
  "in_reply_to_screen_name": null,
  "possibly_sensitive_appealable": false,
  "in_reply_to_status_id": null,
  "id": 828982363610755073,
  "in_reply_to_user_id": null,
  "geo": null
}

Full JSON result for @HilzFuld reply, statuses/show/836964090245447680¶

{
  "id_str": "836964090245447680",
  "place": null,
  "favorite_count": 20,
  "user": {
    "following": false,
    "statuses_count": 164112,
    "id_str": "15930061",
    "listed_count": 2322,
    "name": "Hillel Fuld",
    "default_profile_image": false,
    "is_translator": false,
    "profile_background_color": "010C12",
    "entities": {
      "description": {
        "urls": []
      },
      "url": {
        "urls": [
          {
            "expanded_url": "http://HillelFuld.com",
            "display_url": "HillelFuld.com",
            "url": "https://t.co/1jWZiUGsAw",
            "indices": [
              0,
              23
            ]
          }
        ]
      }
    },
    "translator_type": "none",
    "description": "Co-Founder @ZCastApp; Blogger TechCrunch | Venturebeat | The Next Web | Huffpo | SAI | Mashable; Mentor Google | MSFT; Startup Advisor; Family Man, Carnivore",
    "utc_offset": 7200,
    "profile_image_url_https": "https://pbs.twimg.com/profile_images/804023141290962946/c5OUZUoG_normal.jpg",
    "protected": false,
    "is_translation_enabled": false,
    "friends_count": 1630,
    "notifications": false,
    "lang": "en",
    "url": "https://t.co/1jWZiUGsAw",
    "profile_sidebar_fill_color": "252429",
    "profile_use_background_image": true,
    "geo_enabled": true,
    "favourites_count": 4692,
    "location": "Israel",
    "created_at": "Thu Aug 21 11:21:02 +0000 2008",
    "profile_sidebar_border_color": "181A1E",
    "has_extended_profile": true,
    "followers_count": 31984,
    "profile_background_image_url_https": "https://pbs.twimg.com/profile_background_images/614764377/e7odsyn7mzlhbi33yvtl.jpeg",
    "screen_name": "HilzFuld",
    "time_zone": "Jerusalem",
    "contributors_enabled": false,
    "profile_image_url": "http://pbs.twimg.com/profile_images/804023141290962946/c5OUZUoG_normal.jpg",
    "profile_banner_url": "https://pbs.twimg.com/profile_banners/15930061/1373391596",
    "profile_background_tile": false,
    "follow_request_sent": false,
    "id": 15930061,
    "profile_link_color": "2FC2EF",
    "profile_background_image_url": "http://pbs.twimg.com/profile_background_images/614764377/e7odsyn7mzlhbi33yvtl.jpeg",
    "profile_text_color": "666666",
    "default_profile": false,
    "verified": true
  },
  "coordinates": null,
  "entities": {
    "symbols": [],
    "urls": [],
    "hashtags": [],
    "user_mentions": [
      {
        "screen_name": "jack",
        "indices": [
          0,
          5
        ],
        "id": 12,
        "id_str": "12",
        "name": "jack"
      }
    ]
  },
  "in_reply_to_user_id_str": "12",
  "favorited": false,
  "text": "@jack How can someone who's been on the platform from pretty much day one give you serious feedback privately?",
  "contributors": null,
  "is_quote_status": false,
  "lang": "en",
  "source": "<a href=\"http://tapbots.com/tweetbot\" rel=\"nofollow\">Tweetbot for i\u039fS</a>",
  "retweeted": false,
  "created_at": "Wed Mar 01 15:39:23 +0000 2017",
  "truncated": false,
  "in_reply_to_status_id_str": "836959185220710402",
  "retweet_count": 2,
  "in_reply_to_screen_name": "jack",
  "in_reply_to_status_id": 836959185220710402,
  "id": 836964090245447680,
  "in_reply_to_user_id": 12,
  "geo": null
}