Respiratory Disease Surveillance in England

A portfolio of public health data science projects for UKHSA Field Services, demonstrating R skills in surveillance, outbreak detection, immunisation, statistical analysis, and environmental epidemiology.

View the Live Dashboard →

---

📁 Project Structure

#	Project	Focus	Output
1	Routine Surveillance	Respiratory mortality trends	HTML
2	Outbreak Investigation	C. difficile detection	HTML
3	Local Authority Request	MMR coverage in Leicester	HTML
4	Statistical Comparison	North‑South hypothesis test	HTML
5	Team Briefing	One‑page synthesis	HTML
6	Environmental Hazards	Air pollution and respiratory mortality	HTML
7	Interactive Dashboard	Exploratory data tool	Dashboard

---

📊 Data Sources

Source	Description
Fingertips	Public health indicators (respiratory mortality, MMR coverage)
UKHSA Data Dashboard	Infectious disease surveillance (C. difficile)
ONS	Population estimates and boundaries
NHS Digital	Hospital Episode Statistics
UK-AIR (DEFRA)	Air pollution data (PM2.5)

---

🔑 Key Findings

• North‑South divide: Northern regions have 14.1 more respiratory deaths per 100,000 – a 48% higher rate (p < 0.001, Cohen's d = 2.4)
• Outbreak signal: East Midlands showed 80% of weeks exceeding C. difficile threshold
• MMR decline: Leicester fell from 95.8% coverage (2013) to 88.4% (2024)
• Sustained inequality: The North‑South gap has persisted for over two decades
• Environmental link: Positive association between PM2.5 and mortality in 2024 (r = 0.57), persisting after deprivation adjustment

---

🛠️ Technical Notes

Aspect	Detail
R version	4.5.2
Key packages	tidyverse, plotly, flexdashboard, leaflet, fingertipsR, ukhsadatR, effectsize
Colour schemes	Viridis (colourblind‑safe)
Interactive elements	plotly for charts, leaflet for maps, DT for tables

---

📋 Skills Summary – Evidence for Data Science Roles

This portfolio demonstrates a range of technical and professional skills applicable to epidemiology, public health data science, and data curation roles.

Technical Skills

Skill	Tools & Methods	Evidence in Portfolio
Data acquisition	fingertipsR, ukhsadatR, openair, API integration, data caching	Projects 1, 2, 6
Data wrangling	tidyverse, string manipulation, fuzzy matching, joins, pivot operations	All projects
Statistical analysis	t‑test, Cohen's d, Poisson thresholds, variance ratio, correlation, partial correlation	Projects 4, 6
Visualisation	ggplot2, plotly, leaflet, heatmaps, colourblind‑safe palettes (viridis)	All projects
Dashboard development	flexdashboard, DT tables, interactive elements	Project 7
Reproducibility	Quarto, Git, GitHub, data caching, project structure	Entire portfolio
Data curation	Reusable pipelines, data validation, quality checks, documentation	Projects 1, 6 (data dictionaries, curation reports)

Professional Competencies

Competency	Evidence
Analytical skills	Welch t‑test, Cohen's d, Poisson thresholds, correlation analysis, p‑value interpretation
Report writing	One‑page team briefing, Quarto documents with clear interpretation for non‑technical audiences
Data governance	Use of official APIs, data caching, reproducible methods, no redistribution of raw data
Communication	Plain‑language briefing, stakeholder‑focused outputs (local authority request), interactive dashboard
Quality assurance	Data validation checks, missing value handling, outlier detection, reproducible workflows
Equality and inclusion	North‑South inequality analysis, health disparities focus, colourblind‑safe visualisations, accessible outputs
Independent work	Structured project with clear phases; self‑directed stakeholder requests

How This Portfolio Maps to Specific Roles

UKHSA Epidemiology and Information Analyst (EO)

Criterion	Evidence
Degree or equivalent	PhD (Imperial) + full R workflow with data caching, version control, and reproducible methods
Analytical skills, basic statistical tests	Welch t‑test, Cohen's d, Poisson thresholds, descriptive statistics
Experience of report writing	One‑page team briefing, Quarto documents
Maintaining/analysing data using common software	R, tidyverse, ggplot2, plotly, leaflet, flexdashboard
NHS / epidemiology experience	Real UKHSA data (Fingertips, UKHSA dashboard), respiratory mortality, C. difficile outbreaks, MMR coverage
Equality of opportunity	North‑South inequality analysis, health disparities focus, colourblind‑safe visualisations

HDR UK / Data Curation Roles

Criterion	Evidence
Data curation pipelines	Reusable data processing scripts, validation functions, data dictionaries
Data quality assessment	Missing value handling, duplicate detection, range checks, quality reports
Documentation	Data dictionaries, curation reports, reproducible workflows
Open science	Public GitHub repository, version control, documented code
Working with messy data	Handling of API data, inconsistent formats, missing values

General Data Science Roles

Criterion	Evidence
Programming (R)	Extensive use of tidyverse, plotly, leaflet, flexdashboard
Statistical reasoning	Hypothesis testing, effect sizes, correlation, p‑value interpretation
Data visualisation	Interactive plots, maps, heatmaps, colourblind‑safe palettes
Reproducible research	Quarto documents, Git version control, cached data
Communication	Clear interpretation for technical and non‑technical audiences

How to Verify

• Live dashboard: marcoSorbona.github.io/ukhsa-respiratory-and-outbreak-surveillance/06_day6_dashboard.html
• GitHub repository: github.com/MarcoSorbona/ukhsa-respiratory-and-outbreak-surveillance
• All code is public, documented, and reproducible

---

🔁 Reproducibility

All analyses are fully reproducible:

1. Clone this repository
2. Run 00_setup.qmd once to download and cache all data
3. Render any day's .qmd file or the dashboard

# Example: render the dashboard
quarto::quarto_render("06_day6_dashboard.Rmd")

---

Documentation

• Information Governance
• Data Quality Prevention
• Equality, Diversity and Inclusion

---

📬 Contact

Marco Sorbona, PhD
LinkedIn | GitHub

---

📄 License

Code: MIT © Marco Sorbona
Data: Open Government Licence v3.0 (via Fingertips, UKHSA, ONS, NHS Digital)