This repository contains web-based visualizations to view, explore, and gain insights into the ISW Web Scrape and NLP Enrichment dataset.
The ISW dataset contains publications from the Institute of the Study of War, including their metadata, full text, and extracted keywords, places, and people names. Visualizations were created using various fields from this dataset to present publication information in an accessible, interactive way.
-
Co-Occurrence Matrix: View how often two people are mentioned in the same ISW publication. See which people are often associated with each other in reporting. Live Demo
-
Force-directed Network: View a network of the most mentioned people, where links are defined by mentions in the same ISW publication. Better understand the key people that publications report on. Live Demo
-
World Map: View a map of the world that highlights how many ISW publications mention each country. Learn where the majority of publications report on. Live Demo
-
Timeline: View how often a person was mentioned in ISW publications on a monthly basis. Track trends of increasing or decreasing reporting on a person. Live Demo
-
Word Bubble: View how frequently individual words appear in ISW publications. See prominent terms used repeatedly throughout publications. Live Demo
A copy of the ISW Web Scrape and NLP Enrichment dataset is located in the data directory. Some visualizations utilize this dataset directly, while others rely on newly formatted datasets derived from it.
Create updated versions of these datasets by running the format_data.py Python script. This script loads the ISW dataset from isw_products.json, builds new datasets used by the visualizations, and writes them out in JSON format. Versions of these files are included in the data directory.
# Go to data dir
cd data
# Create JSON data files
python3 format_data.py