Or, use: http://lockerproject.org/ and export everything locally without it passing through someone else's web server (as well as keep your local copy of the data completely current for as long as you decide to keep your locker running.)
Nice use of the twitter bootstrap, I just started using it as well to kick off new mini projects. Just gave twitter a whirl, looking forward to seeing the export.
Also saw you are using gauges, how do you like the reporting?
gaug.es is really nice for tracking social engagement and responding to traffic trends in realtime... It doesn't retain a ton of data for long-term analytics, so if you care about that sort of thing, it should probably be used in conjunction with another service.
It's against Twitters TOS to provide a service that can present a users tweets as a downloadable dump. This new restriction was brought in earlier this year.
The Google+ one was added by user request. Yes, you could use Google Takeout to get the data as well.
I am not storing any of the data or even your OAuth tokens after the dump has been completed. As an implementation detail, your email address may remain in a log file for up to five days.
This is an awesome project, if not because it makes the concept of exporting your data from these sites very transparent, and approachable for the average person.
If you're comfortable with releasing your code, it would be awesome to have some sort of "offline" mode where users could opt in to downloading your scripts and running them locally. That way you'd be affording ease of use and allaying anyone's security concerns. Of course releasing code opens up a whole other can of worms (maintainability, license), but until then I wouldn't feel comfortable running my exports through your sites, as much as I support the idea. Others might feel differently though, in which case the site has a ton of utility.
The exporting scripts are pretty trivial just using the APIs each of these services expose. The hardest part is setting up the OAuth bits, which require a web server to host a callback endpoint. This makes it non-trivial for the average person to run the scripts and becomes a nightmare for support.
That said, I might open source it at some point just so that people can audit it and verify that I'm not doing anything malicious.
I've tried to construct a social graph from my own data using the Twitter API as a side project, but it seems that the rate limiting of 350 API calls/hr really kills what I can do. Anyone have experience or insight into an approach to this?
Full disclosure: I work on this!