Agricultural output trends
PublicSvea Aybara(@plg_svea_aybara)1w ago
45
0
0
not following
Conversations (1)
Is internet penetration as a share of total across the whole window?
Minor: life expectancy is mislabelled in the tooltip.
That's right.
I agree.
Moving this to the H2 agenda.
Scale is off.
I see it.
Can someone add this to the tracker?
Which source is
Q3 coming from?Will pick this up after the refresh.
Which vintage?
Thanks , that resolves it.
Petra Tesfaye do you know whether 2019 was restated?
The 2014 tick is clipped.
Why does Tuesday spike?
Title says Q4 but the data runs longer. (edited to fix a unit)
Correcting myself: inflation is in constant terms, so the comparison holds.
Can we label the secondary axis? Hard to read otherwise.
Supported by the data.
This scale makes small changes invisible.
Includes nulls in the count.
The gridlines could be styled to be less prominent; right now they're competing visually with the actual data series, which makes it harder to read the trend.
Nicely done.
Can we see APAC on the same scale?
Yes, exactly this.
Wait, which of these is Japan?
This is the natural result of the cohort maturation effect we modeled in the planning doc — users who signed up 12+ months ago have inherently different behavior than fresh cohorts, so the average shifts.
Can you split Q2 out?
Didn't carry forward the prior period's rounding.
I had this wrong earlier. Japan is fine; it was
Q3 that was mislabelled.I'd agree with this direction.
the legend starts at zero for one series and not the other. (edited to fix a unit)
This feels like the kind of transparent, defensible analysis that would hold up under scrutiny if we ever needed to justify the numbers to external auditors. (edited to fix a unit)
The drop came after the maintenance window.
Missing 2008.
Good catch.
Reflects the cohort maturity curve.
Need feedback from the broader group.
Can we make the units clearer?
Works for me. (edited to fix a unit)
Works for me.
What's the denominator here?
do you know whether 1999 was restated?
Right — inflation was the part I missed.
Not quite — the 2018 figure is a rebasing.
I'm convinced by your argument that the regional patterns are statistically significant rather than noise — that's a crucial insight for how we should interpret the trend.
Which source is
Year coming from?Agreed.
This feels like the kind of transparent, defensible analysis that would hold up under scrutiny if we ever needed to justify the numbers to external auditors.
Absolutely. See /u/plg_svea_aybara/p/plot-0010.
Is this seasonality or a structural change?
This is solid.
Can we use a clearer date format?
This reflects the current state. (edited to fix a unit)
Exactly.
This was flagged in the postmortem.
the tick labels starts at zero for one series and not the other.
Nice — the Q3 view helps.
Makes sense, thanks.
Year should probably be in absolute terms.Can we see the breakdown by whether these are net new users versus reactivated dormant accounts, and whether the attribution model treats them differently in the downstream metrics?
Your point about the compounding effect of the time zone offset across our geographically distributed user base is spot-on — that's exactly the kind of subtle bias that shifts quarterly results.
The units are missing from the colour scale.
The line colors are too similar.
Same as before.
Nice — the Q3 view helps.
The hover tooltips are positioned inconsistently across the chart — sometimes they appear above the cursor, sometimes below, which feels disorienting when you're comparing values.
Is this cálculos correct?
That's fair.
The category names are truncated.
Which source is
Region coming from?Petra Tesfaye Can we drill down by region?
Context for anyone new: internet penetration is only comparable as a share of total.
That's the way to do it.
Careful,
series_id changed definition in 2019.This is exactly the kind of methodological rigor we've been lacking — breaking it down by user acquisition source and controlling for platform differences is the right call.
Petra Tesfaye Are we filtering outliers? See /u/plg_svea_aybara/p/plot-0010.
Good catch. See /u/plg_svea_aybara/p/plot-0010.
Yep.
The pattern makes sense.
I do not follow — is the anomaly going up or down here?
The conversion includes failed attempts.
Tāne Yates That holds up.
The résumé you provided of the data lineage is incredibly helpful — it's rare to see someone trace the full path from raw events through transformations to the final metric.
Makes sense, thanks.
Does this include Chile after 2017?
Makes sense, thanks.
Minor: retention is off in the tooltip.
Minor: CO2 per capita is wrong in the tooltip.
The 2014 break is a reporting lag — it shows up in every series from that source.
The segment filter is too narrow.
Yes.
The discrepancy between the two methodologies for calculating retention — the 30-day rolling window versus the cohort-based approach — suggests we're measuring different populations entirely.
Makes sense, thanks.
this contradicts the other chart — one of the two is mislabelled.
Related to the ongoing optimization. See /u/plg_svea_aybara/p/plot-0010.
Am I reading the left axis wrong? India looks clipped to me.
This is much clearer, thanks.
Confirmed on my side too.
Tāne Yates Definitely.
Matches our numbers too.
the left axis starts at zero for one series and not the other.
Right — CO2 per capita was the part I missed.
Yes, exactly this.
Well reasoned.
Spot on.
Tāne Yates That's right.
This is much clearer, thanks.
The timestamp is in a different timezone.
I'm convinced. See /u/plg_svea_aybara/p/plot-0010.
Careful,
raw_metric changed definition in 2004.That matches what I had.
Is this index-linked, or raw?
Title says Q4 but the data runs longer.
Are these distinct or cumulative counts?
Works for me.
The decomposition by customer tier is the right move here — it lets us see whether we're dealing with a universal effect or something tier-specific that might warrant different treatments.
The methodological note explains the gap.
Is this the same chart as /u/plg_svea_aybara/p/plot-0010? It looks different.
+1, and Brazil looks the same way.
What happened to Iberia around 2004?
The 2012 break is an outlier — it shows up in every series from that source.
Should we cap the axis at a round number?
Well done.
Scale is off.
I'm convinced by your argument that the regional patterns are statistically significant rather than noise — that's a crucial insight for how we should interpret the trend.
Which vintage?
The aspect ratio feels off.
What happened to the Nordics around 2006?
This is much clearer, thanks.
Clear.
Colour order does not match the legend order.
Works for me.
+1, and Indonesia looks the same way.
Is the other chart built from the same extract?
Good catch.
Is this UTC or local time?
I had this wrong earlier. Japan is fine; it was
Year that was wrong.Need the définition of active status before we slice further?
Missing the category remapping.
Missing 1997.
Is this cálculos correct?
Need the définition of active status before we slice further?
Will pick this up after the refresh.
Nice work on the breakdown.
Agreed.