spot check errors

[hwchen]

Victor spot checked 6 profiles: 2 places, 2 counties, 2 states, for 8 indicators.

All were correct except for 3:

• Median Household Income for Austin, TX
• Citizenship for Austin, TX
• Citizenship for New York, NY

I've confirmed the issue for all three: The number that's shown is equivalent to the 1yr est, but the profile says it's source from 5yr. I think that this is probably not a backend issue, since I've checked and the cube numbers match AFF. And since I'm going on vacation, I'll leave the rest of the investigation to you!

Median Household Income for Austin TX:

1yr = 67,755 5yr = 63,717 Victor got these numbers from American Fact Finder, I confirmed that these numbers are the same in the cube.

Citizenship for Austin TX:

1yr = 87.2% 5yr = 87.4% Victor got these numbers from American Fact Finder, I confirmed that these numbers are the same in the cube. (It's actually a calculated percentage, the cube only gives raw population)

Citizenship for New York, NY

1yr = 83.7% 5yr = 83.3% Victor got these numbers from American Fact Finder, I confirmed that these numbers are the same in the cube. (It's actually a calculated percentage, the cube only gives raw population)

[jspeis]

This seems like an issue that's happening on the site side of things, as it's currently using both 1&5 year data but only noting the use of the 5-year in the viz. Ideally, since Austin, TX is a big city this viz would only use 1-year estimates in this case

[davelandry]

fixed on redwood!

this was caused by the parents and similar logiclayer relations, which would (in these cases) return an MSA the has a large overlap with the selected ID, but has a population less than 250,000, which would cause the entire viz query to go to the 5-year estimate. I made both of those relational endpoints only return IDs greater than 250,000 if the selected ID is also big.