We don't often see the +/-1, +/-2 and +/-3 sigma (σ) zones explicitly shown on control charts. Trying to work out if lines are enough or whether grey shading works better.

I've already spent too much time this weekend thinking about statistical process control.

Apologies for the poor resolution of this, but I quite like the way my earlier attempt at a stem-and-leaf plot has now morphed into a pair of 'granular histograms'. This is a comparison of ED fullness values on 39 'bad' days (68% compliance) with 46 'good' (96% compliance) days.

This is a really interesting talk on silos by @gilliantett. I particularly like the bit between 39:00 and 44:00 when she talks specifically about how data can be analysed in ways that make clear the relationship between different silos. youtube.com/watch?v=Avi9xfoJ…

On Reporting Scotland just now, John Swinney looks clueless in response to @GlennBBC's questions about corridor care and winter pressures. No wonder @VP_Scot_RCEM is livid.

Taking inspiration from John Donne's poem To His Mistress Going to Bed for a control chart =IF() statement in Microsoft Excel. As one does. =IF(AND(C2>=H2,C2<=I2),"BETWEEN",IF(C2>I2,"ABOVE",IF(C2<H2,"BELOW","")))

Climbed the Cobbler today. Lots of mist swirling about.

I want to find a way of measuring bed occupancy in a way that's *patient*-centred as opposed to ward-centred. Maybe by measuring the average number of beds occupied in the ward, but just for the duration of a particular patient's stay. Bed occupancy as witnessed by the patient.

I have rashly volunteered to do a talk at the next EdinbR meeting (on 21st October). I'm calling it "For a Few Dollars More", and I'm asking the question: "Can we use the language conventions of R to help develop our thinking about whole system problems?" #rstats