NoSQL_Project

GDELT NoSQL project The aim of this project is to explore relations between countries in 2018 based on specific queries on the GDELT database.

Architecture

Configuration Cassandra

Replication Factor = 3 Write = QUORUM(2) Read = ONE(1)

W + R = RF Eventual consistency

EMR Automation

Scripts to automate the cluster creation: 1 - bootstrap_cassandra.sh : Bootstrap cassandra on cluster creation 2 - cluster_configuration.py : Configurate the cluster (number and types of instances etc) + link to spark and cassandra 3 - create_cluster.sh : Launch to create EMR

Data Loading and preprocessing

For data with events and mentions tables : https://github.com/sarah911/NoSQL_Project/blob/master/2E4E4Q6WY/note.json

For data with gkg table: https://github.com/sarah911/NoSQL_Project/blob/master/2E1J1S7FX/note.json

Queries

Q1: Find the number of articles and events for a triplet ( Data, Country, Language )

Q2: Find events of an actor in the past 6 months

Q3: Find actors with the most negative or positive views based on ( Date, Country, Language )

Q4: Find actors, countries and organizations that divide the most given a date

Q5: The evolution of relations between countries

Part 1: Based on actors names (Table events)

Part 2: Based on actors countries (Table mentions)

Part 3: Based on articles written in a country about another one (Table GKG)

The final presentation