Scaling a Baby Name Database From 500 to 2100 Names: Lessons Learned

BabyNamePick started with about 500 carefully curated names. We're now past 2,100. Here's what we learned scaling a structured dataset while keeping quality high. The Quality vs Quantity Trap It's tempting to bulk-import name lists from public datasets. We tried this early on and quickly reverted. The problem: inconsistent data quality. Origins were wrong, meanings were oversimplified, and gender classifications were outdated. Instead, we add names in curated batches of 20-30, each manually verified for: Accurate origin(s) — many names have multiple cultural roots Nuanced meanings — not just dictionary definitions Current gender usage — some names have shifted over time Popularity scoring — based on recent data, not historical Data Structure Evolution Our initial schema was flat: { name : " Sage " , gender : " unisex " , origin : " latin " , meaning : " wise " } At 2,000+ names, we needed more structure: { name : " Sage " , gender : " unisex " , origin : [ " latin " ], meaning : " Wise