dig into plausibility of use gap seasonality

[bertozzivill]

• are the residuals more or less seasonal than the raw data?
• how do these results compare to those found in the Koenecker paper?
• plot against aggregated versions of covariates (?)

[bertozzivill]

For now: plot with monthly-aggregated survey data

[bertozzivill]

The size of the uncertainty intervals makes this a less important point.