Add WB prevalence of stunting/wasting data/functions

[jm-rivera]

Add functions to download and store:

• Prevalence of Stunting, measured by height for age (modeled estimate, % of children under 5) https://data.worldbank.org/indicator/SH.STA.WAST.ZS

• Prevalence of Wasting https://data.worldbank.org/indicator/SH.STA.WAST.ZS

Compare coverage and latest data to that available from WFP. Report back in comments

[jm-rivera]

For wasting

• WFP has data for 140 countries. No detail on when the observation is from, however.
• WB has data for 169 countries. Some of the data is very old.

70% of the numbers are the same, 8 are only available from WB (generally wealthier countries). The mean absolute difference (for non-equal values) is 1.3pp, the median 1.05, the max is 4.7pp

Recommendation: use WFP data enhanced by WB only in cases where it is missing.

For stunting

• WFP has data for 140 countries. No detail on when the observation is from, however.
• WB has data for 156 countries. Some data is very old.

70% of the numbers are the same, 7 are only available from WB (generally wealthier countries). The mean absolute difference (for non-equal values) is 2.5pp, the median 2.1, the max is 7.2pp

Recommendation: use WFP data enhanced by WB only in cases where it is missing.

[jm-rivera]

Leaving open for a decision on which data to use for product.

[amy-dodd]

I'd prefer to use wasting - as a more generalised indicator of poor nutrition across populations - if the data stands up to it. Thanks

[jm-rivera]

Closing on the assumption that we'll move forward with wasting, as described.