CPS grapher allow for aggregate microdata in moving average about econ_data

[bdecon]

To implement: in the first data collection step, adjust the code in order to return the individual components of the final calculation rather than the final calculation itself. For example for the employment-population ratio, this would be the sum of weights of employed persons and the sum of weights of the population. Previously, I would calculate the monthly epop.

Then, a 12 month moving sum, for example, could be applied to the separate components, and the resultant moving sum columns could be used in the calculation. This would allow the moving average to adjust for changes to the individual components that might be obscured in the final calculation at a monthly level, but exist in the real world. Example of cases where this makes a difference might be a migrant workforce or calculations involving places with major seasonal population differences.

[bdecon]

Still need to figure out a fast implementation of a moving aggregate to handle the wage by quantile calculations (median real hourly earnings, for example).

[bdecon]

Complete