The one about analyzing free form text by a fairly generic and obvious architecture. Also, Lotus had already implemented something very similar some years before.
The other patent is difficult to understand but it seems to be describing some sort of demuxer. Pretty much any mediaplayer, software or hardware, seems to be "in breach" of this patent.
Those are the only two patents that weren't thrown out by ITC, AFAIK.
Here is a list with links to the patents. Which two are you talking about? How are they overly broad and obvious at the time they were applied for.
http://www.engadget.com/2010/03/02/apple-vs-htc-a-patent-breakdown/
How do you know what data Apple collects?
I read their privacy policy. (Well, skimmed it.
Do you think iCloud will be analyzed by Apple in order to get more interesting data on you? I do.
And there you are shifting goalposts. Collecting data and selling it are two very different things. Of course Apple is going to collect data on their own services.