Team Raksha/SRMIST MOZOHACK 19/Nirbhaya App
Nirbhaya App – Push 01

Tasks Covered: Data Pre-processing

• imported COBRA crime datasets ranging from 2008-2019.

• removed redundant features/columns

• removed null/missing entries

• removed outlier values with low frequencies

• split 24 hours into 12 slots (12:00 AM – 2:00 AM, 2:00 AM – 4:00 AM, and so on…)

• labelled each record with a day of year (1-365)

• suitably labelled each record for the above changes

Next Task : create hierarchy of time block, day, neighbourhood respectively, with count attribute.
: pass to a k-means classifier to get 3 distinct labelled categories of crime.
: will use categories obtained to gauge whether an area is dangerous or not.

Nirbhaya App – Push 02

• created hierarchical structure to label according to crime level

• removed lower lying values from dataset to reduce complexity of said structure

Next Task:
-implementing k-means to label crime levels according to general trends.
-use these labels to gauge whether an area is dangerous or not

Nirbhaya App – Push 03

-Used k-means clustering algorithm on unlabelled dataset created, in order to form 3 separate clusters.

-assigned a category value (1,2,3) to each row.

Next Task: build a progressive web-app and use the category wise safety results from the dataset to avoid unsafe areas along the route.