So, I've managed to download an archive of my Twitter feed. Here's some info for anyone that's not done it yet and is interested...
If you've not done it yet, you can access your archive by selecting the 'More' option along the left-hand side of your profile page. You then need to go to 'settings and support, 'settings and privacy', and select 'Download an archive of your data'.
You'll have to verify your account with a password and confirm via email/text. It will say it takes 24hrs for your archive to become available. For me, it took 4-5 days.
When your archive is available, you'll get an email and a Twitter notification to tell you and you'll get a download link. The link gives you a zip file with LOTS of files inside. It is BIG! For me, it was 1.56Gig and my profile has 28.8k posts.
If you're running an organisational account or are a keen tweeter, your file is going to be a lot bigger. So, what's in the box...?
Your archive.html will give you an offline version of your Twitter profile, where you'll get access to your tweets, stats and more.
Interesting things to see include which advertiser lists you're on.
And what Twitter thinks you're interested in which may not be entirely accurate...
So this all works entirely offline, and with a bit of editing, looks like a tweet archive could easily be made available via an organisational website, if you strip out the stats and the links to them.
Next up is the two folders - 'assets' and 'data'. 'assets' is there to make the HTML interface work so you can pretty much ignore that - just keep it with the HTML file if you're going to move it. The 'data' folder is where the good stuff is.
Within the 'data' folder, you'll find a bunch of files and some folders. The folders are where you'll find all your uploaded media, the main one being 'tweets_media'. Both images and video are included and all animated GIFs are stored as MP4s.
The original filenames are replaced unique IDs, so good luck if you're wanting to find anything in particular.
The rest of the files in the folder make up your actual data. Excel? CSV? No - that would be too convenient... *.js files...
Fortunately, the naming is fairly sensible and tweets.js is where your main repository of tweets is found. This file is actually a JSON file in disguise which, while being nonsense to us humans, is easily readable by computers.
In theory, this should mean organisations can move tweet data into a collections management system as digital objects...
Now, these files will need some TLC to get them usable by us mere humans. Open up the file in a text editor and delete the bit I've marked here. Then save it but change the extension to .json
*ACCESS NOTE* I've not included ALT-text in these images as code and filenames is probably not going to play nice with screen readers. If you're a screen reader user, please do DM if you'd like and I can try and explain in a way that works for you.
So now we have a database file but it's not in a format that's readable/usable for most of us. From what I understand, Excel will import this nicely if you have a fairly recent version and you're on a PC (but cavets - more later!).
If you're on a Mac like me, you'll have a bit more work to do. Basically, you'll want to convert the file to a CSV and options to do this seem to be few and far between. Especially if you want to do it offline for data security.
I used EasyDataTransform which is available for both Windows and Mac and has a time-limited free trial. Seeing as you only need to do this once, this was a good option for me. easydatatransform.com
I won't go into too much detail here but here's a screengrab of exporting to CSV...
So now we have our CSV file, we can export to Excel! All hail Excel! Except Excel will need a bit of hand-holding... One of the most important columns is tweet.id
_str (there is also tweet.id
but ignore that one for *reasons*)
While Twitter is alive, we can use this to make links to Tweets on Twitter online.
Unfortunately, if we try and bring this number into Excel, Excel treats it as a number but doesn't play nice with numbers that big, converting to scientific notation and replacing the last four digits with empty, useless zeros.
What I've found works for me is when importing into Excel ('Get Data' from the Data menu) is to set 'Data Type Detection' to 'Do not detect data types':
Et voilà - All your tweets, likes, retweets, etc in a lovely Excel spreadsheet!
Then you can use pivot charts! Totally useless chart of tweets by tweet length anyone?
Number of tweets by handle replying to - I think @Mar Dixon ☕️🎭🗿🏛❤️
and I may be in a bubble! Shout out to anyone else here that's been keeping me company over the years!
The rise and fall of tweeting. There's definitely a story behind these numbers!
How to make yourself feel more liked on Twitter: use a logarithmic Y axis...
All-time top tweet. This kind of stuff is not easy to find in standard Twitter analytics but would be useful for orgs benchmarking against past performance. Also, yay Kylie! twitter.com/soupdragon2000…