extreme values in doi_10.7910_DVN_28557.R

[rhijmans]

@cedricngakou, some of the values in your PR #825 are way outside what is possible. Can you please investigate and address that?

/survey/doi_10.7910_DVN_28557.R
    unknown variables: farmer_age, farmer_education_level, farmer_marital_status, plot_slope, seed_price, market_type
    NA values: yield, yield_marketable
    out of bounds: plot_area (0.5, 404700), yield (0, 7e+06), seed_rate (0.027, 1235483.105), fertilizer_amount (0.48, 66666.667), fwy_residue (0, 7e+06), crop_price (0.04, 1200000), yield_marketable (0, 156494.516)

[cedricngakou]

Issues with this dataset:

• Some units do not have a converter from local to standard units. (This was addressed by filtering the data to retain only rows with a valid converter to the standard unit (kg).)
• Some plot areas are too small compared to the crop weight harvested, resulting in very high yield values (kg/ha).
• There are probably some erroneous values that occurred during the data recording process (especially for crop prices).

[rhijmans]

This is unfortunately common in surveys. The best we can do in most cases is to filter out the obvious errors. But if there are very many then the entire dataset should be discarded --- as other values may be within the plausible range by chance.