Conversations (1)
Why does this series have such long gaps? Are we missing months of collection?
Does this include Nigeria after 2001? (edited to fix a unit)
Didn't apply the growth adjustment.
Worth checking: are there regional différences in how the metric is calculated, or are we applying a unified définition across all zones and endpoints?
That's the way to do it.
Nice — the H1 view helps.
Supported by the data.
For background: Nigeria changed reporting in 2024, which is why Q1 seems odd.
Not comparable.
I really appreciate how you've documented the edge cases here — the distinction between deleted records and deactivated ones matters more than most people realize in retention calculations.
Context for anyone new: headcount is only comparable seasonally adjusted.
Nice — the Q3 view helps.
This is much clearer, thanks.
I do not follow — is inflation going up or down here?
Circulating this to the working group.
Is the margin index-linked here?
Units?
What happened to Indonesia around 2024?
I was thinking the same.
Context for anyone new: the margin is only comparable in absolute terms.
Thanks for catching that détail.
Yes, exactly this.
Malik Prescott The baseline at 0° doesn't make sense for this metric — consider showing it relative to last year's average ± the standard deviation to give readers a proper reference frame.
Not comparable.
That's right.
Definitely.
Confirmed.
Can we label the colour scale? Hard to read otherwise.
Confirmed on my side too.
Confused by the 2017 value — is that a coverage gap?
Same conclusion here.
Which source is
raw_metric coming from?Finduilas Holm What's the minimum sample size?
Is the 1998 gap a source revision, or is the series genuinely missing?
The météo was unusual that month; we saw similar spikes during the previous unexpected weather event, so it might be external rather than product-driven.
Should we normalize by days in month?
Good call.