Compute demographic distribution metrics

[futurechris]

What we have so far:

• [x] Income distribution by FMA
• [x] Household language by FMA
• [x] Households with occupants 65+ by FMA
• [x] Housing tenure (renter/owner) by FMA
• [x] Race by FMA
• [x] Total population by FMA

What we will have soon:

• [x] Health insurance status
• [x] Poverty status
• [x] Receipt of food stamps/SNAP
• [x] Educational attainment
• [x] Geographical mobility

The task here is to look for correlations between this demographic data and our two key fire-related metrics:

• [ ] # of incidents
• [x] average response time

We'll need to consider how to appropriately normalize the data:

• [ ] by population
• [ ] commercial zoned sq footage (esp. for fire incidents)

[kielejocain]

Blocked by #4

Edit: #4 broken up into tasks; now blocked by #38

[sky-t]

Updated a notebook to look at median response time vs various demographic metrics. I think I may have found something interesting (response time may have some relationship w/ socioeconomic status) that we can discuss on Wed and may merit further investigation: https://github.com/hackoregon/emergency-response/blob/analytics/notebooks/census_eda_geo.ipynb