The hotspots and Kernel density surfaces were derived for total accidents as a whole and particularly
for accidents during monsoon and non-monsoon times and incidents near to educational institution and
religious places. The hotspots for each classified types of accidents were calculated using the Getis-Ord
Gi* function followed by the event calculation. The Getis-Ord Gi* statistics identifies spatial clusters of
high values (hot spots) and of low values (cold spots). The output of hotspot analysis tool is GiZScore and
GiPValue for each feature. These values represent the statistical significance of the spatial clustering of
values, given the conceptualization of spatial relationships and the scale of analysis. A high GiZScore and
small GiPValue (probability) for a feature indicates a spatial clustering of high values where as a low
negative GiZScore and small GiPValue indicates a spatial clustering of low values[21,27]. The higher the
GiZScore, the more intense is the clustering. A Z score near zero indicates no apparent spatial clustering.
The GiZScore for unclassified total accident locations varies between -2.093 to 7.781 and the GiPValues
from 0 to 0.998. From this, it is inferred that statistically significant positive GiZScore (high values)
indicates accident hotspots, while statistically significant negative GiZScore (low values) indicates
coldspots. In the case of accidents during monsoon and non-monsoon seasons, the GiZScore and
GiPValues range from 5.556 to -1.250 and 4.286 to -2.171 and 0 to 0.211 and 0.029 respectively. Similar
hotspots Gi* statistics were observed for accidents near educational institutions and religious places. The
GiZScore ranges from 5.753 to -1.552 and 5.748 to -1.384 with GiPValues 0 to 0.120 and 0 to 0.166
respectively for incidents near educational institutions and religious places