Why Data Matters
Good programs start with good data. Before designing interventions, you need to understand: Where is malnutrition concentrated? What are the underlying causes? Who is most vulnerable? What resources exist?
This page catalogs the most useful data sources for nutrition programming. Some are global databases you can access online. Others are country-level surveys you'll need to request. All are more reliable than guessing.
⚠️ Data Limitations
No dataset is perfect. Survey data may be 2-5 years old. Coverage is often incomplete. Methods vary across countries. Use data as a starting point, not gospel. Validate with local knowledge and recent rapid assessments.
What Kind of Data Do You Need?
📚 Major Nutrition Data Sources
These are the most important databases for nutrition program planning:
👥 Demographic & Geographic Data
Population & Demographic Data
- WorldPop: High-resolution population density maps and demographic estimates. Essential for estimating target populations in specific areas.
- UN Population Division: Country population projections and demographic indicators. Useful for medium-term planning.
- National Census Data: Most detailed demographic data available. Contact national statistics offices. Often 10+ years old.
- World Bank DataBank: Development indicators including poverty, education, health infrastructure. Good for contextual analysis.
Geographic & GIS Data
- OpenStreetMap: Free, editable maps with roads, facilities, and administrative boundaries. Variable quality but improving.
- GADM: Database of global administrative boundaries. Download shapefiles for any country free.
- HealthSites.io: Crowdsourced database of health facility locations. Incomplete but useful where official data unavailable.
- NASA FIRMS: Active fire data for tracking agricultural burning, conflict, or displacement. Real-time updates.
🎯 Assessing Data Quality
Not all data is created equal. Before using a dataset for program decisions, assess its quality:
🛡️ Using Data Responsibly
⚠️ Data Can Cause Harm
Data about vulnerable populations is sensitive. Malnutrition data can stigmatize communities. Detailed location data can enable targeting by armed groups. Ethnic/religious breakdowns can fuel discrimination.
Principles for Responsible Data Use
1. Privacy Protection
- Never share individual-level data without consent
- Aggregate data to at least district level before sharing publicly
- Remove geographic precision (GPS coordinates) before publication
- Consider whether ethnic/religious breakdowns are necessary or could cause harm
2. Context & Interpretation
- Always provide context: "20% wasting" needs comparison to emergency thresholds
- Explain limitations: sample size, geographic coverage, data age
- Avoid ranking or shaming: "District X has highest malnutrition" can stigmatize
- Present trends, not just snapshots: improving or worsening?
3. Avoid Misuse
- Don't use data to justify predetermined conclusions
- Don't cherry-pick indicators that support your narrative
- Don't compare incomparable data (different methods, seasons, populations)
- Don't claim precision you don't have (confidence intervals matter)
4. Community Ownership
- Share findings with communities surveyed, not just donors
- Use accessible language and formats, not just technical reports
- Allow communities to respond to or contest findings
- Credit local data collectors and knowledge holders
🔍 When Data Doesn't Exist
Perfect data rarely exists for the exact question you need answered. What to do when there's a gap:
📋 Options for Filling Data Gaps
- Use proxy indicators: If you don't have wasting data, poverty and food insecurity can indicate high-risk areas
- Extrapolate from nearby areas: If neighboring districts are similar, their data may roughly apply
- Rapid assessments: Small surveys (100-200 households) can give rough estimates quickly
- Key informant interviews: Local health workers and leaders know which areas are worst affected
- Program monitoring data: Screening data from existing programs reveals trends even if not representative
- Satellite/remote sensing data: Crop conditions, rainfall, vegetation indices correlate with malnutrition
✅ Best Practice: Triangulate
Don't rely on one data source. Use multiple sources and methods. If survey data, key informant interviews, and health facility records all point to the same districts as high-need, you can be more confident. Contradictory data signals the need for better information before committing resources.
🚀 Next Steps
If you need data for program planning:
- Start with DHS or MICS for your country—check recency
- Look for SMART surveys or nutrition cluster data if in humanitarian context
- Download administrative boundaries and facility locations from HDX or OpenStreetMap
- Supplement with local knowledge and recent rapid assessments
- Document data sources, limitations, and assumptions for stakeholders
If you're collecting new data:
- Use standardized methods (SMART, WHO) for comparability
- Get ethical approval if required (university, ministry of health)
- Train enumerators properly—bad data worse than no data
- Build in data quality checks during collection
- Share findings responsibly, protecting privacy
- Use our Data Collection Protocol Builder