The way I use Twitter is to download tweets to a local database, including images/videos/unshortened URL and then I view the data in my own UI from local DB. It circumvents the tweet/account removals done by users or Twitter. Also I can do any kind of search/data processing I want. I guess I'm not alone. You can't hide your data once they are made public. It's interesting what is possible once you start treating web…
How many petabytes does something like this take up?
I didn't optimize for space at all, twitter's json is quite wasteful.