This is not an answer.
I got bored, so, using commoncrawl.org data dated 26 Sep 2018 (but not all collected at the same time, of course), I managed to extract a list of 30,815 .gov domains in the "gov-domains.txt" file at http://test.barrycarter.info/20191107/
A larger list of all 69,051,164 known to commoncrawl.org on the same date is at the same URL in ...
The DIY way would be to to compile the lists of websites with a bit of internet searching:
and then from each page source scrape social media accounts
<div class="usa-footer-secondary_section usa-footer-big-secondary-section">