I have 475 followers as of Dec. 13, 2022, and I want to express my gratitude to each and every one of you for following me. This is a little gift to some of you in return for providing an engaging experience in the Twittersphere. If you are not one of the users already on this page and would like me to create one for you, or if for whatever reason you prefer not to have your tower displayed on this site, please DM me on Twitter.
Twitter timeline towers visualize the 3200 most recent tweets of a specific user's Twitter activity using an approach which resembles a stack of spoked wheels, each wheel being the user's daily tweet/retweet activity. Learn more about the history of this method. This is somewhat of an arbitrary limit given that all of a users tweets are still accessible even if they are older than the most recent 3200, but it is a very real limit.
I used the code on my github site for scraping the most recent 3200 tweets or whatever they tweeted if it was less than 3200. I scraped in 100 tweet increments to keep my rate below what might be throttled. I was able to scrape all of their tweets in less than 48 hours. This resulted in about 10,000 files each containing 100 tweets, or one million tweets, still well within my two million tweet monthly limit.
I sorted the 475 users by followers_count, i.e. how many followers they have. I then generated timeline towers for the top 30. These are shown below, with links to download the ANTz executables for mac and pc.
There are three different executables depending on your OS and existence of features/bugs. They are:
"antz": Mac OS X version, not as many features as the Windows version
"antz.exe": Old Windows version (fewer features, fewer bugs)
"ANTz-XR.exe": Newer Windows version (more features, more bugs)
If you get an error when running either of the Windows versions about a missing DLL, this will hopefully be resolved by installing the Visual C++ redistribution with one or both of the executables found here.
There are two key files, "antz0001node.csv" and "antz0001tag.csv", containing the tweet data and visualization parameters. For the mac and older pc version, these files are located in:
/usr/csv/
For the newer pc version, they are located in:
/usr/antz0001/csv/
If you have any problems running ANTz on your computer, download the latest release and replace the two default data files included in the latest release with the same files in your timeline tower download.
Double-click the app for your computer to run it. Hit the number 1 key to load the data. If it doesn't load, click and drag the mouse for a second and try again. When the data loads you should see the tower with a central axis. That axis is selected by default. Hit the "i" key to display the text associated with the selected object. Then hit the up arrow to jump to the tweet branch level and hit the "i" key again. Click and drag with the left mouse button in an open area to rotate, and with both left and right buttons down to zoom. Select one of the colored spokes (green, white, or gray) and hit the "i" key to display the timestamp of the tweet. Hit the "u" key and that specific tweet should open on Twitter in a web browser. Hit tab to jump to the next object in the level. You can see the level of an object at the bottom of the viewing window when you select it. For more details, view the documentation.
ThatEricAlper has the most followers of all users who follow me, ~878,900. Eric is a very real person but this account is automated. It tweets nearly every 30 minutes, which is 48 tweets per day, so this tower consists of 63 days of tweeting. All tweets are original, and the number of 'likes' are in a similar range. It is interesting that it's a bot. In fact, they don't get much more bot-like than this. But that's not a bad thing. Twitter bots are a useful practical tool for making repeated tasks automated.
Download the Twitter User Timeline Tower for 'ThatEricAlper'.realTuckFrumper has ~214,500 followers. This tower has 8 spoked wheels or 8 days of data. The account also appears to be bot-like. It tweets original tweets throughout the 24-hour cycle, although there appears to be a lull in tweeting at roughly the same time each day. Either this person does not sleep through the night, or it's a bot. Most tweets are original. The keywords 'trump' (red) and 'biden' (blue) appear occasionally. Tweets seem to get a consistent number of 'likes'.
Download the Twitter User Timeline Tower for 'realTuckFrumper'.Stop_Trump20 has ~175,700 followers. Trump was sort of stopped, so this account still seems somewhat active, at least one tweet a day. Tweets are a mix of types, mostly original or reply. Many tweets include a high perentage of special or uppercase characters.
Download the Twitter User Timeline Tower for 'Stop_Trump20'.oXxRowanxXo has ~170,100 followers. They are most likely a real person who gets a good night's sleep. This dataset spans the past year of 2022 almost to the day (Dec. 13). Tweets get a considerable number of 'likes', keywords 'trump' and 'biden' are found occastionally. This account seemed to change four times over the course of several weeks or months from original tweets (lots of green) to reply tweets (mostly white).
Download the Twitter User Timeline Tower for 'oXxRowanxXo'.1drcole has ~158,100 followers. This dataset spans from Oct. 31 to Dec. 13, 2022. They seem like a real person who tweets during half the day and very little during the other half except for one original tweet at 10am GMT every day to #Resisters. This one tweet is likely automated.
Download the Twitter User Timeline Tower for '1drcole'.BradReason has ~154,800 followers. This dataset spans Nov. 21, 2019 to Dec. 12, 2022. There was alot of tweeting during 2020 but waned after the election. There are a variety of tweet types and characteristics, and most tweets are posted during half that day, indicating a real person.
Download the Twitter User Timeline Tower for 'BradReason'.ResisterForever has ~146,600 followers. This dataset spans Jul. 5 to Dec. 12, 2022. Tweeting seems very consistent during that time except for a conspicuous gap during Nov. 20 to 30. A tweet on Nov. 23 indicates why.
Download the Twitter User Timeline Tower for 'ResisterForever'.1JaySC has ~147,000 followers. This dataset spans Aug. 9 to Dec. 12, 2022. They tweet mostly replies (white spokes) distributed throughout 1/2 to 2/3 of the day. There was a daily original tweet regarding 'Democratic meetups' at around, but not exactly, 10am GMT which switched to 11am GMT on Nov. 5. These tweets have similar, in fact nearly identical, characteristics.
Download the Twitter User Timeline Tower for '1JaySC'.smc429 has ~134,500 followers. This dataset spans Dec. 4 to 12, 2022 suggesting many tweets per day. There are daily gaps between 8am and 3pm GMT suggesting this user gets their needed sleep and draws fairly distinct lines between periods of tweeting and not tweeting. This tower has some resemblances to a non-English tweeter due to all of the obvious magenta glyphs representing a high percentage of special characters. However, these are often related to how they did on Wordle that day.
Download the Twitter User Timeline Tower for 'smc429'.DalithSteiger has ~119,500 followers. This dataset spans Jan. 23, 2019 to Dec. 12, 2022. They tweeted actively for most of 2019 and then waned a bit since then. Many tweets include numerous hashtags. 2019 tweets were mostly original, then diversified somewhat since then. Tweeting at the same time each day for numerous days in a row appears to be a common behavior but the time of day varies and the trends are relatively brief.
Download the Twitter User Timeline Tower for 'DalithSteiger'.