Seriously though, 365/7=52.14 + plus leap yrs, causes it periodically
A data scientist's nightmare.
▶️2015 prev one, avgs tricky
▶️3-4 day wks at Xmas/NY disrupt anyway
▶️C19's made normal reporting higher
A perfect storm.
1/3
Plus, we start fresh yr, so best to switch to ONS wks31-30 winter reporting.
But, which avg/stdev gully to use?
🥇byDeath: excess is tiny, PHE use this, but @JuliaHB1's just tried it on using this
🥈byPub: overplays, e.g. @toadmeister style dishonesty
Oh, the dilemma.
2/3
Honesty, everytime.
Facts:
▶️PHE daily C19 nos 1.3k⏫over ONS&PHE wks52&53 reports due to xmas/NY
▶️So est ONS wk1: 6.8k C19
▶️Today's daily PHE nos make ONS wk2 est: 8k+ C19
So I project short dashed lines for ONS wks1&2.
Why do people still see this as a game?
3/3
• • •
Missing some Tweet in this thread? You can try to
force a refresh