Breast Cancer Now Generations Study – Aggregated Data Dashboard

A browser-based dashboard for exploring pre-aggregated, privacy-safe summary statistics from the Breast Cancer Now Generations Study. No individual-level data is used or stored — all values are group-level summaries with small-cell suppression applied.

Live dashboard: https://uk-generations-study.github.io/Aggregated-Data-Dashboard/

---

Overview

This dashboard is designed for sharing aggregated study statistics without disclosing individual participant records. It loads a single JSON file (aggregated_data.json) containing pre-computed summary statistics, and renders interactive charts and tables entirely in the browser.

Privacy protections built in

• No individual records — the JSON file contains only means, medians, quartiles, histograms, and frequency counts
• Small-cell suppression — any count < 5 is suppressed and displayed as <5
• Suppressed strata — any subgroup with fewer than 5 participants is omitted entirely
• No data transmission — all rendering happens locally in the browser

---

Quick Start

Online (GitHub Pages)

Open the dashboard directly — no installation needed:

https://uk-generations-study.github.io/Aggregated-Data-Dashboard/

The aggregated data file is loaded automatically from the same repository.

Local use (with auto-refresh)

If you need to update aggregated_data.json and preview changes locally:

./start.sh          # opens browser at http://localhost:8080/
./start.sh 9090     # custom port

start.sh starts a local HTTP server so the dashboard can fetch the JSON file (required — browsers block fetch() on file:// URLs).

---

Dashboard Tabs

Tab	What it shows
Introduction	Study background, data provenance, privacy notes
Overview	Participant count, completeness, variable groups
Explore	Per-variable charts (histogram, boxplot, bar) and summary stats for the whole cohort
Missingness	Proportion missing and sentinel-NA by variable
Stratified	Compare any variable's distribution across subgroups (e.g. by menopausal status)
Descriptive	Summary statistics table, optionally stratified, with CSV export

---

Repository Contents

├── index.html                 ← Main dashboard (open this in a browser)
├── aggregated_data.json       ← Pre-aggregated summary statistics (no individual records)
├── aggregate_data.py          ← Python script that generates aggregated_data.json
├── start.sh                   ← Local HTTP server launcher (for local preview)
├── logo-g.png                 ← Study logo (dark background)
├── logo-white.png             ← Study logo (light background)
└── README.md                  ← This file

---

Updating the Aggregated Data

The aggregated_data.json file is generated centrally from individual-level data using:

python3 aggregate_data.py input_data.json --min-cell 5

Argument	Default	Description
input_data.json	(required)	Individual-level data file
--min-cell	5	Suppress counts below this threshold

Once regenerated, replace aggregated_data.json in the repository. The dashboard will load the new file automatically (no code changes needed).

---

Suppression Rules

Level	Rule
Frequency table cell	Count < 5 → displayed as <5
Stratified subgroup	n < 5 → entire stratum omitted
Histogram bin	Count < 5 → bin excluded from chart; note shown on chart

---

Requirements

• Dashboard: any modern browser; internet connection for initial Chart.js CDN load
• aggregate_data.py: Python 3.7+ (standard library only)
• start.sh: Python 3 (for local HTTP server)

---

Breast Cancer Now Generations Study — thegenerationsstudy.co.uk