From Nutrition to Newborn: Unlocking Maternal Health Insights with Decomposition Tree
Maternal health is a cornerstone of public healthcare, yet complications during pregnancy and childbirth continue to affect both mothers and infants globally. To better understand these intricate health dynamics in Maternal Health Data Analytics Project—used SQL for data processing and Power BI, specifically the Decomposition Tree, for impactful visualization and storytelling.
Project Overview
This project used Power BI dashboards and SQL queries to explore how different factors, especially maternal nutrition and fat distribution, affect pregnancy outcomes. A key feature was the Decomposition Tree, which helped visualize how individual health variables cascade into delivery and newborn results.
Initial participants: 272 pregnant individuals
Lost to Follow-Up: 61 participants (22%)
Final Analyzed: 211 participants
Dataset
This project incorporated a wide array of variables:
Demographics: Age, ethnicity, reproductive age
Nutrition: Daily intake of calories and key nutrients
Lifestyle: Alcohol, Tobacco, Drug usage
Lab Results: HIV, Hepatitis C, Syphilis, Hypertension, Diabetes
Fat Assessment: Preperitoneal combined fat, Skinfold thickness, Subcutaneous fat
Anthropometry: BMI, waist-to-hip ratios
Labor & Delivery: Delivery type, meconium in labor
Fetal Outcomes: Birth weight, APGAR score, neonatal complications
Tools
SQL
Filtered and cleaned the dataset
Created segments for comparative analysis
Joined multiple tables to build a consolidated dataset
Power BI
Designed dashboards to visualize trends
Created a Decomposition Tree to highlight causal pathways
Identified high-risk combinations and key intervention points
The Decomposition Tree: Tracing the Journey from Diet to Delivery
The Decomposition Tree in Power BI became the storytelling heart of the project. It allowed me to trace a clear path—from maternal nutritional choices to biomarkers like maternal age, through to fat composition, waist hip ratio, hypertension, chronic disease and finally to labor complications and newborn health.
Example Insight:
“Participants with high carb intake showed high fat, obesity , leading to increased visceral fat, which in turn correlated with higher cesarean rates and low birth weight for newborn.”
Why the Decomposition Tree Matters in Maternal Health
Far from being a static chart, the Decomposition Tree is a dynamic decision-support tool that:
Clarifies Complex Relationships
Tracks how upstream factors like diet influence downstream health events.
Enables Personalized Insights
Highlights specific risk pathways for each participant profile.
Supports Evidence-Based Interventions
Helps clinicians target the right stage: nutrition, diagnostics, or monitoring.
Reveals Hidden Risk Clusters
Combines variables (e.g., periumbilical fat, chronic disease) that increase complications.
Interactive Exploration
Allows users to drill into each layer of data and extract precise insights.
Story
The Decomposition Tree acted as a clinical storyboard, showing how seemingly simple lifestyle choices can echo across a pregnancy timeline, ultimately affecting maternal and newborn health.
This project reaffirmed that data is not just about numbers—it's a powerful narrative tool that enables early interventions, customized care, and better outcomes.
Note : Link for dataset:
Link to learn about decomposition tree step-by-step :
