Health Data Scientist Roadmap: From Excel to Global Health Impact in 12 Months – Nic

Why Health Data Science Is Different

Health data science is not just about models—it’s about patients, facilities, financing, and policy.

Your analyses influence:
- Who gets enrolled into a program
- Which facilities receive resources
- Whether a donor continues funding a life-saving intervention
That means:
- You must respect ethics, equity, and context
- You need both statistical rigor and domain understanding

The 4 Pillars of Health Data Science

Data Foundations
- Data types: patient-level, facility-level, claims, survey, registry
- Common formats: CSV, REDCap exports, Kobo/ODK, DHIS2
- Data quality: duplicates, missingness, inconsistent IDs
Stats & Methods
- Descriptive stats, confidence intervals, regression
- Survival analysis for time-to-event outcomes
- Longitudinal analysis for repeated measures
Tools
- R (tidyverse, survival, ggplot2)
- SQL for querying large tables
- Quarto/R Markdown for reproducible reports
Communication
- One-page briefs for program leads
- Visual dashboards for non-technical stakeholders
- Clean, annotated code for other analysts

12-Month Roadmap (While Working or Studying)

Months 1–3: Strengthen Foundations

Excel + basic statistics
Learn R basics:
- Import, clean, transform, visualize
Build 2–3 small projects:
- Vaccination coverage trends
- Facility readiness scores

Months 4–6: Health-Focused Analysis

Learn:
- Regression (linear, logistic)
- Survival analysis for outcomes like time-to-default
- Intro to causal diagrams (DAGs)
Projects:
- Malaria incidence trend analysis
- Health worker density vs outcomes

Months 7–9: Reproducible Research & Dashboards

Learn:
- Quarto for automated reports
- Shiny or basic dashboards (or Power BI if your team uses it)
Build:
- A reproducible health facility dashboard
- A quarterly report pipeline (data → R → HTML/PDF)

Months 10–12: Real-World Projects

Volunteer:
- Support a university department, NGO, or clinic with data cleaning and simple dashboards
Build:
- 2–3 end-to-end case studies you can show employers

Portfolio Ideas for Health Data Scientists

You can stand out with a portfolio that includes:

A dashboard showing hypertension control rates across facilities
A survival curve analysis for time-to-default among patients
A simulation exploring how health financing changes affect out-of-pocket costs

Each project should include:

Data description and limitations
Methods, clearly explained
2–4 key charts
3–5 actionable recommendations

Where to Look for Your First Role

Research Assistant / Data Analyst roles in:
- Universities (schools of public health, epidemiology, biostatistics)
- NGOs running health programs
- Monitoring & Evaluation teams
Keywords to search:
- “Health data analyst”
- “Biostatistics assistant”
- “Monitoring & Evaluation analyst”

You don’t need to start in a “Data Scientist” title. If you own the data pipeline and deliver insights that change decisions, your title will catch up.

--- title: "Health Data Scientist Roadmap: From Excel to Global Health Impact in 12 Months" subtitle: "A practical, beginner-friendly path to become a health data scientist even if you're outside the US & Europe" author: "Nichodemus Amollo" date: "2025-10-27" categories: [Health Data, Data Science, Career, Roadmap] image: "/img/blog-health.svg" --- ## Why Health Data Science Is Different Health data science is not just about models—it’s about **patients, facilities, financing, and policy**. - Your analyses influence: - Who gets enrolled into a program - Which facilities receive resources - Whether a donor continues funding a life-saving intervention - That means: - You must respect **ethics, equity, and context** - You need both **statistical rigor** and **domain understanding** --- ## The 4 Pillars of Health Data Science 1. **Data Foundations** - Data types: patient-level, facility-level, claims, survey, registry - Common formats: CSV, REDCap exports, Kobo/ODK, DHIS2 - Data quality: duplicates, missingness, inconsistent IDs 2. **Stats & Methods** - Descriptive stats, confidence intervals, regression - Survival analysis for time-to-event outcomes - Longitudinal analysis for repeated measures 3. **Tools** - R (tidyverse, survival, ggplot2) - SQL for querying large tables - Quarto/R Markdown for reproducible reports 4. **Communication** - One-page briefs for program leads - Visual dashboards for non-technical stakeholders - Clean, annotated code for other analysts --- ## 12-Month Roadmap (While Working or Studying) ### Months 1–3: Strengthen Foundations - Excel + basic statistics - Learn R basics: - Import, clean, transform, visualize - Build 2–3 small projects: - Vaccination coverage trends - Facility readiness scores ### Months 4–6: Health-Focused Analysis - Learn: - Regression (linear, logistic) - Survival analysis for outcomes like time-to-default - Intro to causal diagrams (DAGs) - Projects: - Malaria incidence trend analysis - Health worker density vs outcomes ### Months 7–9: Reproducible Research & Dashboards - Learn: - Quarto for automated reports - Shiny or basic dashboards (or Power BI if your team uses it) - Build: - A reproducible health facility dashboard - A quarterly report pipeline (data → R → HTML/PDF) ### Months 10–12: Real-World Projects - Volunteer: - Support a university department, NGO, or clinic with data cleaning and simple dashboards - Build: - 2–3 end-to-end case studies you can show employers --- ## Portfolio Ideas for Health Data Scientists You can stand out with a portfolio that includes: - A dashboard showing **hypertension control rates** across facilities - A survival curve analysis for **time-to-default** among patients - A simulation exploring how **health financing changes** affect out-of-pocket costs Each project should include: - Data description and limitations - Methods, clearly explained - 2–4 key charts - 3–5 actionable recommendations --- ## Where to Look for Your First Role - Research Assistant / Data Analyst roles in: - Universities (schools of public health, epidemiology, biostatistics) - NGOs running health programs - Monitoring & Evaluation teams - Keywords to search: - “Health data analyst” - “Biostatistics assistant” - “Monitoring & Evaluation analyst” You don’t need to start in a “Data Scientist” title. If you **own the data pipeline and deliver insights that change decisions**, your title will catch up.