Health outcomes form three clear patterns.
Countries fall into high-pressure, transitioning, and strong-outcome groups when maternal mortality, life expectancy, and direction of change are considered together.
FOR PEOPLE WHO WANT THE STORY FIRST
What can health
data tell us about
where care is needed?
This project makes a complex global health analysis understandable: the question it asks, the patterns it finds, and why those patterns matter to everyday lives.
Start with the research question ↓THE PLAIN-LANGUAGE RESEARCH STORYTECHNICAL DETAILS BELOW ↓
01 · THE RESEARCH QUESTION
WE ASKED
Can health data reveal meaningful groups of countries—and help us understand where health gaps are largest?
IN PLAIN LANGUAGE
Yes. Instead of treating every country as a single score, the analysis looks for countries that share similar health outcomes.
That gives a clearer picture of the global health gap: who faces the greatest pressure, who is making a difficult transition, and who has strong outcomes to learn from.
02 · WHAT THE ANALYSIS FOUND
Three findings that
make the picture clearer.
These are not labels for people or nations. They are practical patterns in the data that invite more focused questions and action.
Progress is not the same as equity.
A country can improve on one indicator while still carrying a large health gap. Looking at both indicators makes that easier to see.
Similar patterns support better questions.
Comparing countries with similar health profiles helps focus attention on practical actions rather than one-size-fits-all solutions.
THE RESULTS, IN NUMBERS
The health gap is
large—and visible.
These numbers show the scale of the differences found across the country groups. They make the findings concrete, without reducing people's lives to a statistic.
countries analysed
Countries with enough health information to be included in the analysis.
face the highest pressure
On average, this group has a maternal mortality rate of 311 and life expectancy of 65 years.
have strong health outcomes
On average, this group has a maternal mortality rate of 8 and life expectancy of 81 years.
are in the middle group
These countries show mixed outcomes—important places to look for what helps progress continue.
EXAMPLE COUNTRIES IN EACH GROUP
What the groups
look like in practice.
These are examples, not a complete list. Countries in the same group have similar overall health patterns in this analysis; they are not identical.
| Health pattern | Example countries | What it means |
|---|---|---|
| Highest pressure | Afghanistan · Chad · Nigeria · Somalia | Higher maternal mortality and shorter life expectancy. |
| Transitioning outcomes | Egypt · Kazakhstan · Malaysia · Morocco · Thailand | Mixed outcomes and uneven progress. |
| Strong health outcomes | Australia · Japan · Norway · Spain · Uruguay | Lower maternal mortality and longer life expectancy. |
03 · WHY IT MATTERS
Better questions
can lead to better care.
Health numbers can feel abstract. In real life, they speak to whether a person can safely give birth, whether families can access care, and whether communities can expect longer, healthier lives.
This analysis does not prescribe a single answer. It helps make the gaps visible so decisions can be more informed, more local, and more humane.
04 · HOW THE RESEARCH WAS DONE
From health records
to a clearer story.
The process is designed to be simple to explain, even though it uses data-science methods behind the scenes.
Gather
Use internationally comparable health records for maternal mortality and life expectancy.
Compare
Look at each country as a health profile, not as one number in isolation.
Group
Use the shared patterns to form three clear, interpretable country groups.
Interpret
Translate the patterns into questions health planners, researchers, and the public can use.
Two health indicators
Maternal mortality and life expectancy are selected as complementary measures of health outcomes.
Outcome + direction
Each country is represented by current indicator values plus recent direction of change. Maternal mortality is log-transformed before scaling.
K-Means is selected
K-Means and Ward hierarchical clustering are evaluated on the same standardised features. K-Means has the stronger silhouette score: 0.393 versus 0.381.
Time-aware forecasting
Short-horizon approaches are tested by withholding the last observation per country. Persistence outperforms the tested linear trends for both indicators.
Open notebooks
The repository contains executed notebooks, source links, requirements, and a refresh script for the analysis.
Pattern, not cause
Clustering describes similarity. It cannot establish causality or replace country context, policy knowledge, or health expertise.