Checkpoint 4

Team: the dapper squirrels

Proposed Visualizations

In this section, we verified our hypothesis by using Spark GraphX and Google Colab. We proposed two graph themes:

1. Making nodes of officers and victims by their income, race, locations, and even unsupervised machine learning models to learn the cluster and see if there is a potential connection between officers and victims.

2. Network dynamics of co-accused in each cohort can be interesting. The analytics can be done with the following:

   1. Make use of Triangle Count Algorithms for each cohort.
   2. Make use of the Page Rank Algorithm to find the most connected officer in all cohorts.
   3. How many CRs that officers have and how many co-accused for each cohort.
   4. Compare the top k largest cohort of police officers in high and low socio-economy status.

How to View Analysis

To review our Google colab notebook, please visit this site: https://colab.research.google.com/drive/1YRSb7iiFOVTxTrqlc6p9hdpZka9nRZf9?usp=sharing
An un-interactive version is also available.

Please check every steps mentioned in the notebook for setup. Run the necessary programs and download the necessary jar files.