The 3rd NHS-R Community Conference wrapped up a week ago last Friday, and was a hugely enjoyable event. There was a pre-conference training week, during which I was able to attend Syliva Canelon’s excellent xaringan workshops. They were fantastic.

Then the second week was the main conference week, consisting of plenary talks in the morning, and lightning talks/ workshops in the afternoon.

This was an entirely remote / virtual event – and it worked brilliantly.

When we first started planning it, and the suggestion of a fortnight long programme of events was made, I was concerned that attendees would run out of steam. But having missed the 2019 conference (I’m still gutted about that), I was very pleased that I could attend from the comfort of my own home without having to worry about travel & accomodation and the resulting impact at home of me not being around.

Virtual conferences are hard though – for both the presenters / instructors, and attendees. It’s hard for presenters to ‘read the room’ and gague how well their content is going down. Attendees are being hit with content left right and centre and it’s hard to get time to reflect.

The fact that everyone was buzzing on Friday, and Anastasiia said she was sad to have to stop the broadcast, as it meant the conference was over, speaks volumes.

And by the way, everyone needs to give Anastasiia a huge amount of credit and plaudits for pulling the whole thing together.

I was on the organising committee ( as I was last year – and the fact that I am willing to be part of an organising committee once, let alone twice, also says a lot about how much I love NHS-R), but Anastasiia was the driving force (along with Zoe, Tom and Emma) that made it happen.

At the first NHS-R event, there were about 100 of us. This year we had over 1000 registrants, and at least 600 attending. Presentations and workshop materials are on YouTube and there is enough material there to keep us going for a while.

I wasn’t able to attend everything, and so I don’t want to call these my highlights, because there is still so much I need to catch up on.

However – Mohammed’s opening address was great, a very laid back introduction to the week.

Ben Goldacre delivered a highly influential talk.

I know it was influential, because finally someone else at work has asked to read the ‘Bringing NHS Analytics into the 21st Century’ paper, which I was only too happy to forward on.

Ben didn’t have slides, he talks to camera for the duration and he is very compelling viewing.

Zoe Turner’s “Why do I need to learn R when I can use SQL?” lightning talk was also very timely, and contained some nice examples of where R makes life easier than the equivalent SQL gymnastics.

Chris ‘Manly’ Mainey delivered an excellent talk, as he always does. Seriously, go watch it.

I also enjoyed ‘For love or money? How to generate high revenue from a team of R developers’ by Lisa Clark. Not because I’m interested in generating tons of revenue for anyone, but because of the messages like give developers the tools they need to do their jobs.

On Wednesday there were also lots of really interesting lightning talks.

I would really have enjoyed ‘How to solve surgery waiting times: R, Drake and Operational Analytics for a data-driven future’ and ‘Hosting R-Shiny Server on Ubuntu Behind NGINX’ even more if I wasn’t too busy being a nervous wreck, as I was also delivering one.

I gave a brief overview of runcharter, patientcounter and spccharter (the latter still needs a proper introductory blog post, but I still have some work to do on it).

The common thread through these packages is that they are mainly powered by data.table.

I sometimes see people apologising for sharing R code that isn’t ‘tidy’ as in is not pure tidyverse.

I made the point that it is not wrong to use base or data.table instead, if it works for you.

Not everything has to be tidyverse, although, if you want to be 100% in that domain, then that’s fine too.

It doesn’t always have to be tidyverse

I made the point by stating that some things ARE wrong e.g. pineapple on a pizza, but using data.table isn’t.

I expected to get some heat from that from hardcore tidyverse fans but instead it was the pineapple afficionados getting all up in my business on the chat / Twitter.

(There was a poll the next day and common sense prevailed. Don’t be giving me any pineapple on my pizza).

Anyway, from the low point (me) we swiftly moved to the the high point of the day in Alberto Cairo. What can you say? Brilliant, just go watch it.

Thursday – I missed the morning sessions, but the evening was the RStudio crew, and they delivered, big time.

Friday morning was great, everyone who presented did a brilliant job : Colin Gillespie from Jumping Rivers re (Code) Quality, Tom & Zoe discussing the NHS-R Slack group, Fiona and Emma from the Health Foundation (aside from awesome R code, Fiona has also created the coolest hex sticker yet for Monstr).

Richard Wilson had a really interesting slot comparing R dashboards with some traditional BI alternatives, and it was a shame that sound problems caused that to have to finish ahead of time. I’d have liked to have seen that one in full. Maybe soon we can get a slot for Richard to try again.

Jamie-Leigh Chapman did a great session on integrating R with Qlik Sense.

Then to cap it all off we had a great talk by Federicca Coco and John Burn-Murdoch.

Chris Beeley rounded it all off nicely, before Anastasiia reluctantly turned off the broadcast.

It seemed to all go very well for participants, there was certainly a good buzz on Twitter.

I’m really pleased to be part of this community, I’m even more pleased that it’s continuing to grow, and although in terms of line up and quality, it feels like this will be a hard act to follow, I’m already expecting that next year, it will be even better.

In the meantime, all the videos are on the Youtube channel here, the workshops are here and course materials on the NHS-R community github here.