Weeknotes 2023-05-01
Bank holiday week this week, so only 3 working days to talk about.
Analytics conference
Definitely the highlight of the week, I joined (and spoke at) the first NUH Analytics and Information Conference. This was a great event, and a chance for us to talk as analysts across NUH, and not merely within the individual groups we operate in day to day.
It was attended by some senior folk from the trust, and the takeaways for me were that clinicians are doing really good research using the future tools that we as "vanilla" analysts need to be using more (R, python, git).
Mark S, our deputy medical director and key person in the development of Nervecentre as our central IT system also spoke. NUH is the first hospital I have worked in, and I had not appreciated how uniquely "good" our colleagues feel our information systems are. For what it's worth, the rainbows are not quite so easily visible from where I currently sit, but clearly I do need to go and connect with parts of the business who are reaping benefits, and find out how they are doing it.
My talk was on the need for our analysts to start working in a more open and collaborative way, using R python and git, which will allow us to share analytical methods and projects in a way that has not so far been possible. Too often I find myself in a position that I need to understand the inner workings of a Qliksense app (what is the source data, how has data cleaning been done, what is the analytical method, how is length of stay calculated?), but I can't because the inner workings are in a "black box" that only the original developer has access to. Qlik's collaboration tools are poor, and my talk proposed that that fact is slowing our progress as a trust.
I'm looking forward to helping keep that conversation going, as we collectively arrive at a better way.
External questions
We recieved a list of followup questions to answer this week, following a visit from important external colleagues last week. Answering questions like these is incredibly important, but it's no less frustrating to loose the opportunity to spend time on other progress.
We are moving forwards, though.
RAP to the rescue
Last year we put a lot of work (about 4 months on and off) into generating a detailed report using finance data as a main source. To cut a long story short, it was disused after 5 or so months in use. That was frustrating, but in truth I wasn't terribly worried - we have plenty else to get on with.
Fast forward to this week, and I had reason to check how robust the code was. I was asked why it was disused, so I pulled a fresh set of data, ran the script, and out popped the updated report. A couple of one-line tweaks to correct {ggplot} and {tidyr} version updates, and the report is back in commission. Now the work on (re) adoption re-starts, but it was hugely satisfying to have it up and running again in under 15 minutes.
More of the same, please (hopefully without the adoption gaps).
Some annual leave and another bank holiday next week, so that will be another short one.
Cheers!