Incidents by Month
Cause of Incidents
Incidents by Route
Years of Experience of Drivers with Incidents
Below is a correlation matrix including Amount of Incidents, Previous Experience, Experience with GSI, and Total Experience. There is nothing significant here except a correlation between Previous Experience and Total Experience, which is not surprising.
| timer.Incidents.to.Date |
1.0000000 |
-0.0839753 |
-0.3037740 |
-0.1841833 |
| timer.Prev..Exp |
-0.0839753 |
1.0000000 |
0.5553425 |
0.9403188 |
| timer.GSI |
-0.3037740 |
0.5553425 |
1.0000000 |
0.8051955 |
| timer.Total.Exp |
-0.1841833 |
0.9403188 |
0.8051955 |
1.0000000 |
From the model below we also see that nothing is significant. This could be due to the scarce amount of data.
lmout <- lm(timer$Incidents.to.Date~timer$Prev..Exp + timer$GSI + timer$Total.Exp)
summary(lmout)
##
## Call:
## lm(formula = timer$Incidents.to.Date ~ timer$Prev..Exp + timer$GSI +
## timer$Total.Exp)
##
## Residuals:
## Min 1Q Median 3Q Max
## -1.3280 -0.8767 -0.5223 0.3787 2.9730
##
## Coefficients: (1 not defined because of singularities)
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.39544 0.44067 5.436 2.16e-05 ***
## timer$Prev..Exp 0.04098 0.08315 0.493 0.627
## timer$GSI -0.21673 0.14489 -1.496 0.150
## timer$Total.Exp NA NA NA NA
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1.305 on 21 degrees of freedom
## Multiple R-squared: 0.1027, Adjusted R-squared: 0.0172
## F-statistic: 1.201 on 2 and 21 DF, p-value: 0.3207
One thing that is hurting this analysis is the lack of data on Previous experience. You can see below that 44.7% of the entries have an NA which means we have no data.
sum(is.na(timer$Prev..Exp))
## [1] 34
length(timer$Prev..Exp)
## [1] 76
34/76
## [1] 0.4473684
Hits were significant when looking at cause of incidents.
The Res and Com departments were also significantly different than other departments.