HackEbola (with Data!)

INSIGHT: preliminary correlations between economic/demographic factors and epidemics metrics

The goal was to correlate economic/demographic data we have with 5 epidemics metrics (total_deaths, total_cases, death_rate_per_person (total_death/population),
case_rate_per_person(total_cases/population), death_rate_per_case(total_death/total_cases)) across regions.
The legend for the economic factors is here: http://www.qdatum.io/public-sources.
The original data was from the "merged_data.csv" file from the "processed data" discussion post (https://www.dropbox.com/sh/ief4x9yshmd6619/AADJgB49a_xQcwl9aYynv2Pua?dl=0).

These are correlations with coefficients with abs. value greater than 0.4 ("strong"):

death_rate_per_person:
edyr_male 0.422613

case_rate_per_person:
edyr 0.402917
edyr_male 0.433035

death_rate_per_case:
age6069 -0.412699

total_deaths:
edyr 0.555448
edyr_fem 0.614050
edyr_male 0.509781
urban 0.576233
qual_floor 0.449826
flush_toilet 0.489299
age2029 0.504105
bad_floor -0.484920
bad_toilet -0.418081
age09 -0.475637

total_cases:
edyr 0.547483
edyr_fem 0.624584
edyr_male 0.492244
urban 0.582566
qual_floor 0.455691
flush_toilet 0.471868
age2029 0.508171
bad_floor -0.480577
bad_toilet -0.432699
age09 -0.477522

The anticorrelation between rural setting (bad_floor and bad_toilet) and total cases/deaths might indicate some kind of reporting bias.