I am not sure if the title accurately reflects my problem but I essentially would like to aggregate a set of metrics of similar nature that comes from different data sources into a single metric. Say we are measuring MetricA from SourceA, MetricB from SourceB,...and MetricN from SourceN sampled over the same general population. There is no guarantee that the union of samples will cover all in the population at least once. With varying sampling rates and frequencies across sources, I would like to come to some level of normalization to sum it up at the top and track that over time.
I am not sure what the best way is to come up with weights though. Has any of you come across with a similar problem or perhaps some papers that found reasonable approaches to avoid/minimize bias?
Many thanks for your suggestions!
