I can only speak for my school, but in my experience, I would say the formality level is between college and business casual for just normal classes. Most people don't wear sweatpants anymore, but a clean pair of jeans are okay. I have seen people in leggings but not often, and typically it's clear that they're going to or from the gym.
And like someone else said before me, if there is an event like a poster session, people do tend to wear business casual or, if it's something really important, even a level above that.
My school teaches Stata mainly, but here's my impression of the three packages from a year of study:
- Stata - Very easy to pick up with a lot of online tutorials and support, but less common in the "real world"; many teachers say they can read Stata code but don't work in it themselves. Personally, I like Stata a lot. Not free, as wwmmkk said.
- SAS - MANY people here use SAS in their day to day work. Not Apple/mac compatible. Can't speak to how difficult or hard this is to learn.
- R - Many people have commented that they think people are moving towards R in the future. Free. The barrier to entry is high, though - "R makes easy things hard but makes hard things very easy", is how I've heard it described. I definitely intend to take an R programming course in the future.
- Python - A few of the more statistically-oriented people I've met here (PhD students and professors) have mentioned wanting to pick up Python. Don't know much more about it than that.