We are in a very weird situation as first_open data for the past months is decreasing a lot (around 1300 for Pakistan).
On 8th of April the count was 10,593, on 9th of April it was 10,405, on 10th of April it become 9,476 and on 11th of April it become 9,332.
I am running,
SELECT geo.country,event_date,COUNT(*) FROM `myproject.analytics_xxx.events_*`
WHERE
AND (geo.country='Pakistan')
AND (event_name='first_open')
GROUP BY event_date
The problem is that it is showing a big discrepancy, which impacting our business.
We are also seeing around 1500 records with geo.country is empty(without any country set).
Here are the firebase ids with specific query,
As per current data in Big Query you will see,
| Country | Date | Count |
|---|---|---|
| 20220224 | 150 | |
| 20220223 | 966 |
Empty(not-set) country with around 1,200 first_open events.
As per our old data (not available in Big Query anymore) around 1,300 events were reduced .
Look like Firebase for Google Analytics is updating the old data (specially updating February data in April) in Big Query. Is this correct?
As per this thread from Firebase support BigQuery to Data Studio discrepancy for event 'first_open',
Normalization is only done for the last 3 days. But we are facing issue with 2-month-old data.
