Scatter of each variable and crime

For now, I am doing only comunas, to see if it leads anywhere and get the code running. Once I can do it with barrios, the sample sizes will increase.

The color in these scatters is each comuna, each point is a day.

Regression of each variable and crime

Naive Regressions

Here I just regress amount of crime in a day, to the measure of PM2.5 of each model.

Dependent variable:
crime
(1) (2) (3) (4) (5) (6)
mean 0.001***
(0.0002)
median 0.001***
(0.0002)
max 0.001***
(0.0001)
min 0.0003
(0.0003)
q25 0.001***
(0.0003)
q75 0.001***
(0.0002)
Constant 0.086*** 0.089*** 0.087*** 0.105*** 0.095*** 0.087***
(0.005) (0.005) (0.005) (0.003) (0.005) (0.005)
Observations 81,913 81,913 81,913 81,913 81,913 81,913
R2 0.0002 0.0002 0.0003 0.00001 0.0001 0.0002
Adjusted R2 0.0002 0.0002 0.0002 0.00000 0.0001 0.0002
Residual Std. Error (df = 81911) 0.546 0.546 0.546 0.546 0.546 0.546
F Statistic (df = 1; 81911) 19.371*** 17.324*** 21.368*** 1.084 9.882*** 18.189***
Note: p<0.1; p<0.05; p<0.01

Here I do the same, but add a “barrio” fixed effect.

Dependent variable:
crime
(1) (2) (3) (4) (5) (6)
mean 0.0004*
(0.0002)
median 0.0003
(0.0002)
max 0.0004***
(0.0001)
min -0.0001
(0.0003)
q25 0.0001
(0.0003)
q75 0.0003
(0.0002)
Observations 81,913 81,913 81,913 81,913 81,913 81,913
R2 0.00004 0.00002 0.0001 0.00000 0.00000 0.00003
Adjusted R2 -0.002 -0.002 -0.002 -0.002 -0.002 -0.002
F Statistic (df = 1; 81763) 3.067* 1.458 11.242*** 0.022 0.091 2.487
Note: p<0.1; p<0.05; p<0.01