Update for the last week, based on the given data from @coinlocket$
As the attached image shown, the means and medians of weekly distributed merits are nearly 4.8k and 4.4k, respectively.
As always, I strongly suggest to take a look at medians, which are almost true means of distributed merits, rather than means,
Using medians will help us to have more exact, neutral overview on merit distribution without significant effects of outliers (robust to outliers).
More interestingly, the interquartile range (from 25th quartile to 75th quartile) has been from 3978 to 4820.
What does it means?
It presents that 25% of observed weeks have less than 3978 merits distributed over week (below the 25th quartile), and 25% of them have more than 4820 merits distributed over week (above the 75th quartile).
In other words, 50% of those observed weeks have weekly distributed merits ranges from 3978 to 4820; and the median of weekly distributed merit has been somewhere around 4.4k per week.
Data interpretation used statistics in the last row of the attached table.

In addition, weeks which have total distributed merits above 6k should be consider as potential outliers [as red circles shown in the boxplot]

*Notes:
1. Interquartile range: The range from 25th quartile to 75th quartile.
2. Mean +/- sd: Mean +/- Standard deviation
3. Potential outliers: weeks which have total distributed merits outside the two whiskers (above and below).
Above whisker = Q3 + 1.5 IQR = 4820 + 1.5*842 = 6083 (Q3: the 75th quartile)
Below whisker = Q1 - 1.5 IQR = 3978 - 1.5*842 = 2715 (Q1: the 25th quartile).