Methodology question: What population coverage threshold makes sense for calculating aggregate GDP per capita for country groups?

I'm building an open-source GDP per capita visualization tool (https://gdppercapita.fyi) using IMF World Economic Outlook data.

For country groups (G7, EU27, BRICS, World Bank income groups, etc.), I calculate aggregate GDP per capita as:

sum(gdp) / sum(population)

The problem: Some countries have missing data for certain years (Syria, Eritrea, etc.). Rather than showing no data for the entire group, I use a 90% population coverage threshold — if countries representing ≥90% of the group's population have data, I calculate the aggregate.

My questions:

1. Does 90% seem like a reasonable threshold? Too high? Too low?
2. Is this a standard approach in economics research, or is there a better methodology?
3. Should I flag years with incomplete data differently in the visualization?

Any references to academic papers or standard practices would be appreciated.