Some fixes and improvements to `using SMRA with R`

[Moohan]

The main changes here are to fix some of the examples near the bottom to do with cohorts, where the object names (including the database connection) weren't consistent and so the examples were confusing.

I've also made lots of other minor tweaks to the text and code, to improve the code style and to make the document use 'proper' markdown - I think it had been copied from Word and still had some artefacts such as bullet points as symbols. The code has been put into proper code blocks to enable code highlighting.

I haven't tested all of the examples (hopefully if they worked before, excepting the incorrect object names, they will still work!), I also haven't updated any of the content/text, particularly I've not updated anything around the new PWB infrastructure etc,