Recently, The Guardian encouraged their journalists to "regularly delete historical tweets". Read a digest at Nieman's Lab.
Here is a repo and a pyhton script to guide you to accomplish this task.
You would first need to get authorization to use the Twitter API and download your Twitter archive. Then, in the archive folder, find the tweet.js
file and make it a JSON file finding the first line...
window.YTD.tweet.part0 = [ {
and replacing it with
{"data": [ {
... and for last line you need to add an extra "}" to make it a JSON object. Your file's last line will look like this:
} ]}
Feel free to give me your feedback.
I found the original solution here.