[RFC] How to handle multiple observation types for a single parameter in the same year and location

[Bluesy1]

For reference, the issue can be observed in the following image:

Looking at Akron, OH in 2011 as an example, the carbon monoxide values are fairly close, so they could probably just be averaged, but comparing the different values for sulfur dioxide, there's a wide range of values depending on the metric.

What should the preferred order for metrics be for us, and how should we handle this?

@IfYouWantMoney

[theHDarian]

While yes, there is a somewhat significant difference in these values, I don't believe we need such fine precision to do our analysis, so we should average them in the interest of time. if we find that greater precision is indeed required, it should not be a monumental task to revert our earlier decision.

[Bluesy1]

@IfYouWantMoney lets say I do just aggregate by a simple average
Thats going to skew readings like this one (the second parameter under the first location above)

I feel like thats at best suboptimal, at worst causing the chance for incorrect analysis/assumptions of the data

[theHDarian]

instead of taking the mean, we can just take the median. it should prevent serious skewing incidents like you suggest.

[Bluesy1]

Makes sense, should be a good middle ground!