Human Value Index: Visualizing Media Bias in Violent Death

Asif Farhan Khan

AI Automation

Data Scientist

AI Developer

AWS

OpenAI

Python

Arts & Entertainment

Human Value Index

Client Overview

Steven is a practicing artist and social entrepreneur. For the Human Value Index project, Steven envisioned an art installation that would challenge audiences to confront the imbalance in how violent deaths are valued and represented in mainstream Western media. He wanted to move beyond abstract statistics and build a live, evolving system that visually highlights the disparity between global conflict-related deaths and the amount of media attention they receive.

Driven by both a conceptual and critical lens, Steven brought forward a powerful vision: to turn hard data into an engaging, thought-provoking display that would question not only media bias but also society’s perception of human value through visibility. This project reflects his broader mission of using technology and design as tools to reveal hidden patterns and spark conversation.

A data-driven art installation that visualizes the relationship between violent deaths due to conflict and their representation in mainstream media coverage, displayed as a live, dynamic index.

Process Design

The vision was to portray the bias of Western media by quantifying "human value" through the lens of acknowledgment and coverage of violent crimes worldwide. This involved developing a mathematical framework, performing extensive data gathering, and ultimately visualizing results in a compelling, ticker-like display.

Workflow Design & Implementation

Our approach began with iterative sprints. First, we built a blueprint for one country, refining data pipelines, computations, and visualizations before scaling globally.

Each sprint included requirements analysis, research, data sourcing, script development, database modeling, and validation. By the end of each sprint, we had a more refined system.

Understanding underlying requirements

Extensive research & Data Hunting

Script & Database Development

Data Modeling & Visualization

Review & Next Sprint Iteration

Sprint I: Parameter Calculation

We sourced data from UNO, WHO, UCDP, and ACLED databases to derive key parameters (year-wise death counts, global deaths, and population metrics). Additionally, we developed an AI-based News Violence Indicator to quantify media coverage intensity and sentiment.

SELECT 
        COUNT(*) AS FilteredCount 
    FROM 
        db.public.analyzed_headlines 
    WHERE 
        CAST(violence_score AS NUMERIC) > 6.0 
        AND sentiment_score LIKE '-%'
        AND category NOT IN (
            'wellness', 'snapchat', 'royalwedding', 'galleries', 'shopping-uk', 
            'happybodies', 'travel', 'money', 'podcasts', 'shopping-us', 'auhome', 
            'shows', 'shoppinguk', 'aussport', 'weather', 'health', 'debate', 
            'yourmoney', 'coffeebreak', 'shopping-au', 'tvshowbiz', 'wellness-us', 
            'aushowbiz', 'sciencetech', 'property', 'motoring', 'sport', 'wellness-uk'
        ) 
    LIMIT 500;

Scraping Process Extracting Headlines from Articles using NLP

Final Database depicting country wise death toll based on news headlines

Custom Trained model identifying "Violence Score" based on the article headline

Data Sources

Technologies Used

Client Remarks

"Asif is very helpful and creative and I look forward to continuing to work with him on this project." "Asif has continued to work hard in pursuing a solid outcome to this project. It's genuinely exciting to see everything coming together. Really appreciate your perseverance on this!"

Like this project

Posted Nov 22, 2025

Developed a data-driven art installation visualizing media bias in coverage of violent deaths.

Likes

Views