Yeah, with mitmproxy, I've done similar things in the past. Much easier to block the whole site, like orders of magnitude simpler. But if you want a project which will force you to learn basically everything about modern networking and web technologies, go for it.

Network-wide requires a bit of effort. As long as you can install a root certificate to the trusted certificate store on the devices you want to block it on, HTTPS won't throw any errors on their devices. You would need a computer of some variety (RPi-esque should be fine) to run the proxy and a DNS proxy/forwarder to force the devices to talk to the computer when trying to talk to twitter.com, and that DNS server has to be configured as the one your router tells DHCP devices to use. Unless you switch the whole router out and run the reverse proxy there instead, then no DNS stuff to figure out.

But for instance you could make twitter dot com/fucker show a generic error page or something. Deeplinks to specific tweets as well. If they use Twitter apps and things that leverage its API you'd have to spend a bit of time doing that too.