Statistical Analysis Link

Evidence does not change lives until it changes decisions.

One of the greatest lessons I have learned in public health is that the quality of our decisions depends on the quality of the questions we ask—and the methods we use to answer them.

Logistic regression is often introduced as a statistical model for predicting binary outcomes. In reality, it is much more than that.
It is a decision-support tool.

In geriatric research, it helps identify which factors truly increase or decrease the likelihood of important health outcomes such as falls, frailty, hospital readmissions, cognitive decline, malnutrition, medication-related complications, and mortality. More importantly, it enables clinicians, epidemiologists, researchers, and policymakers to prioritize interventions where they will have the greatest impact.

The infographic below was developed to simplify one of the most widely used analytical methods in epidemiology and clinical research. It moves beyond formulas to demonstrate how logistic regression translates raw data into actionable evidence for healthier ageing.

It covers:

• What logistic regression is and when to use it

• The mathematical model and probability calculations

• Odds Ratios (OR) and their interpretation

• Step-by-step worked examples in geriatric studies

• Confidence Intervals and p-values

• Common analytical pitfalls

• Practical applications in epidemiology, surveillance, and evidence-based decision-making

As populations continue to age globally, the ability to identify risk factors before adverse outcomes occur will become increasingly important. Predictive analytics, when applied responsibly, can help health systems shift from reactive care to proactive prevention.

Data alone cannot improve health.

Evidence interpreted correctly can.

I hope this resource supports students, researchers, clinicians, epidemiologists, data scientists, Monitoring & Evaluation professionals, and public health practitioners seeking to strengthen evidence-based practice.

I would be interested to hear from professionals across different sectors:

In your experience, what is the most frequently misunderstood aspect of logistic regression—and how can we improve statistical literacy among health professionals?

:

Choosing the Right Model for the Right Problem📈🤖

Regression is one of the most fundamental concepts in Machine Learning—but not every regression algorithm is built for the same type of data.

The visualization below provides a simple comparison of some of the most widely used regression techniques and how they fit different patterns in data.

🔹 Here's what each algorithm is best suited for:

✅𝐋𝐢𝐧𝐞𝐚𝐫 𝐑𝐞𝐠𝐫𝐞𝐬𝐬𝐢𝐨𝐧 – Ideal when the relationship between variables is linear and easy to interpret.

✅ 𝐒𝐭𝐨𝐜𝐡𝐚𝐬𝐭𝐢𝐜 𝐆𝐫𝐚𝐝𝐢𝐞𝐧𝐭 𝐃𝐞𝐬𝐜𝐞𝐧𝐭 (𝐒𝐆𝐃) 𝐑𝐞𝐠𝐫𝐞𝐬𝐬𝐢𝐨𝐧 – Efficient for training on large-scale datasets with faster optimization.

✅ 𝐃𝐞𝐜𝐢𝐬𝐢𝐨𝐧 𝐓𝐫𝐞𝐞 𝐑𝐞𝐠𝐫𝐞𝐬𝐬𝐢𝐨𝐧 – Captures non-linear relationships by splitting data into meaningful decision rules.

✅ 𝐑𝐚𝐧𝐝𝐨𝐦 𝐅𝐨𝐫𝐞𝐬𝐭 𝐑𝐞𝐠𝐫𝐞𝐬𝐬𝐢𝐨𝐧 – Combines multiple decision trees to improve accuracy and reduce overfitting.

✅ 𝐊-𝐍𝐞𝐚𝐫𝐞𝐬𝐭 𝐍𝐞𝐢𝐠𝐡𝐛𝐨𝐫𝐬 (𝐊𝐍𝐍) 𝐑𝐞𝐠𝐫𝐞𝐬𝐬𝐢𝐨𝐧 – Predicts values based on the closest neighboring data points.

✅ 𝐍𝐞𝐮𝐫𝐚𝐥 𝐍𝐞𝐭𝐰𝐨𝐫𝐤 𝐑𝐞𝐠𝐫𝐞𝐬𝐬𝐢𝐨𝐧 – Excels at modeling highly complex and non-linear relationships in large datasets.

✅ 𝐗𝐆𝐁𝐨𝐨𝐬𝐭 𝐑𝐞𝐠𝐫𝐞𝐬𝐬𝐢𝐨𝐧 – One of the most powerful ensemble algorithms, widely used in machine learning competitions and production systems.

✅𝐒𝐮𝐩𝐩𝐨𝐫𝐭 𝐕𝐞𝐜𝐭𝐨𝐫 𝐑𝐞𝐠𝐫𝐞𝐬𝐬𝐢𝐨𝐧 (𝐒𝐕𝐑) – Effective for handling non-linear data while maintaining strong generalization.

✅ 𝐏𝐨𝐥𝐲𝐧𝐨𝐦𝐢𝐚𝐥 𝐑𝐞𝐠𝐫𝐞𝐬𝐬𝐢𝐨𝐧 – Extends linear regression to model curved relationships between variables.

If your city is planting fruit trees, share this with your urban planner because waiting 10 years for the first harvest is completely optional.

Farmers have known this for centuries. When they find an exceptional fruit tree — highly productive, outstanding quality — they don't just admire it.

They clone it with mud, straw, and a plastic wrap. It's called air layering.

And the results are hard to ignore.
🌿 Make a few cuts on a mature branch of your best tree
🌿 Wrap it in wet mud and straw — a living bandage
🌿 Cover with plastic film and wait ~30 days
🌿 Roots emerge from the wound
🌿 Cut below the roots. Plant it. Done.

No nursery. No guesswork. No decade of waiting.
Because the new tree grew from a mature branch, it skips the entire 5–10 year juvenile phase.

It fruits in year one. Same quality as the parent. Shorter height. Easier to harvest in tight urban spaces.

Cities investing in edible urban forests are making the right call. But if we're planting at scale, we should be cloning the best performers — not gambling on random seedlings.

This is exactly why we built TREESABLE. Most urban greening decisions are still made on instinct — plant something, hope for the best, never actually measure if it's working.

Treesable quantifies the real ecosystem benefits trees deliver, so cities and developers can stop guessing and start designing with the evidence in hand.

Extreme Poverty by Country (1992–2026) Global extreme poverty—measured by the World Bank's absolute international poverty line of $2.15 per day—has shifted dramatically between 1992 and 2026. While the global rate plummeted from over 35% in the early 1990s to single digits, this progress has been unequal and highly concentrated.

Historical Trends (1992–2020)

* China's Historic Drop: 🇨🇳 went from having the world's largest extreme poverty in 1992 to nearly eradicating it by 2020.

* India's Ascent: Over 400 million Indians were lifted out of extreme poverty during this period.

* Regional Stagnation: Poverty increasingly localized in due to fast and regional .

* The COVID Shock: The 2020 pandemic caused the largest global poverty spike since , briefly reversing decades of progress.

Current Global Distribution (2026)

In 2026, extreme poverty is highly localized in Sub Saharan Africa and conflict-affected regions. The highest absolute numbers of people in extreme poverty reside in:

* 🇳🇬
* 🇨🇩 (DRC)
* 🇮🇳

Countries with the Highest Extreme Poverty Rates

By percentage of the population living below the international poverty line, the nations facing the most severe extreme poverty include:

*
*
* 🇧🇮
* 🇸🇴
* 🇾🇪

Key Drivers and Outliers

* Sub Saharan Africa: Houses roughly 60% of the world's extreme poor, with 18 of the 20 poorest countries globally.

* Conflict-Driven Reversals: Nations like 🇸🇩 have seen massive poverty spikes.

* Asian Success: and the have almost entirely eradicated , while continues to reduce its extreme poverty headcount despite pockets of high inequality.

🚨 Trump may have accidentally made the strongest argument against American football.

“We have to come up with another name for the NFL because clearly, it's not football.”

“It doesn't even make sense when you think about it.”

😂

And honestly...

he's not wrong.

In actual football, people use their feet.

In the NFL, players spend most of the game:
• throwing the ball
• carrying the ball
• tackling people

The foot shows up occasionally and somehow gets the entire sport named after it.

🚨 What's funny is that the rest of the world has been saying this for decades.

To billions of people:

Football = football.

NFL = hand-egg.

So after years of Americans arguing that everyone else is wrong...

the President of the United States just walked into the debate and said:

"Yeah, this name doesn't make much sense."

🚨 Somewhere in Europe, South America, Africa, and Asia...

millions of football fans are nodding in agreement for the first time ever.

Trump may have started many arguments.

But this might be the first one where half the planet immediately agrees with him.

One of the biggest productivity killers in the workplace is not a lack of talent, resources, or commitment. It is a lack of clarity.

When employees are given unclear job descriptions, conflicting instructions, and competing priorities, they often find themselves pulled in multiple directions at once. One manager wants one thing, another wants something different, and the employee is left trying to satisfy everyone while excelling at nothing. The result?

✔ Confusion increases.
✔ Accountability decreases.
✔ Performance suffers.
✔ Frustration grows.
✔ Teamwork breaks down.

Just like an individual cannot successfully pursue two different paths at the same time, employees cannot consistently deliver excellent results when they are unsure of what success actually looks like.

Great organizations understand that clarity drives performance. People perform best when they know what is expected of them, what their priorities are, and how their contribution fits into the bigger picture.

Growth requires clarity. Success requires commitment. Excellence requires focus.

As leaders, one of our most important responsibilities is not simply assigning work—it is providing direction. When people know exactly where they should focus their time, energy, and effort, they become more confident, productive, accountable, and effective.

A confused workforce will always struggle to achieve outstanding results. When priorities are clear, performance improves. When focus is sharpened, excellence becomes possible.

.

📊 Data vs. Evidence: Understanding the Difference

Many people use data and evidence interchangeably, but they are not the same.

🔹 Data are raw facts, figures, and observations collected from surveys, reports, monitoring systems, or research activities.
🔹 Evidence is what emerges when data is analyzed, interpreted, and placed in context to answer a question or support a decision.

Think of it this way:

➡️ Data tells us WHAT happened.
➡️ Evidence tells us WHY it happened and WHAT it means.

In Monitoring & Evaluation, research, and decision-making, collecting data is only the first step. Real impact comes from transforming that data into evidence that can guide policies, improve programs, and drive meaningful change.

💡 Data alone informs. Evidence influences action.