I really enjoy the environment of the DTC. We have recently started a weekly event called “Coffee, Cake and Science Talk”. Its main purpose is to get DTC students together and have a chat about their work (or unrelated things…) while enjoying some nice cake.
Last time we entered a really fruitful and encouraging discussion about making you research truly reproducible and coding practices. One website that got mentioned is Software Carpentry, which is exactly about writing code to make your research reproducible. I found this website particularly useful, since recently I had to go on a parameter goose chase for a model that I needed to reimplement.
The website mostly talks about things that actually are (or should be) common sense, but setting up repositories for your code or keeping exact track of software versions you use is a bit annoying. So at least I find it quite useful to be reminded of all these things. After looking at the website I set up a Git repository on Bitbucket, which gives you unlimited private repositories. Normally I would always be on the open source side, but if you’re working on a project that involves unpublished stuff, you might rather choose to not entirely open it to the public straight away.