May 2026
Drives Beat Drops at Every Level. 304K Shots Prove It.
I analyzed 304,000 pickleball shots and built an interactive dashboard in an afternoon. Here's what the data says — and what it proves about building with AI.
The Debate
Every pickleball player has an opinion on the third shot: drive it hard or drop it soft? Coaches teach the drop. YouTube teaches the drive. Rec players argue about it between games. Pros hedge.
The problem is that most of the debate runs on feel, not data. “Drops work better at higher levels” is repeated so often it sounds true. But nobody checks.
I wanted to check. PKLMart publishes a shot-level dataset — 304,000 shots across 935 games and 10 skill tiers, from 2.5 rec through Senior Pro. Enough data to settle the question. So I built a dashboard to explore it.
The Data
The headline finding: drives outperform drops at every single skill tier.Not just at rec. Not just in men's doubles. At every level from 2.5 to Senior Pro.
The gap narrows at Pro level — but it never flips. The conventional wisdom that drops “take over” at higher levels doesn't hold up in 304K shots of data.
- Higher-rated players sustain longer rallies — not shorter
- Forced errors account for ~50% of all point endings
- Court zone patterns shift dramatically above the 4.5 skill tier
- Server advantage varies by rally length in predictable ways
All six analyses are interactive — filter by skill tier, compare across levels, drill into the specifics. The dashboard makes the data explorable, not just reportable.
The Build
I built the entire thing in an afternoon. One person, no data team. I pointed Claude at the PKLMart CSV and piloted it through the full stack: ETL pipeline, statistical aggregation, and a six-chart interactive dashboard.
This is what I call FYBY — For You, By You. The person closest to the domain question is the person building the tool to answer it. AI handles the syntax. You bring the judgment about what to measure and what matters.
Ingest
Python script parses the raw CSV, validates fields, and computes per-tier aggregations across 40K+ points.
Analyze
Win rates, rally distributions, zone breakdowns, point endings — all pre-computed into a JSON artifact.
Visualize
Next.js + Recharts dashboard with tier filtering, responsive layout, and auto-generated OG images.
Same Architecture. Different Data.
The pickleball domain made this fun to build. But the architecture is the same pattern I use for clients: raw data in, pipeline through, interactive dashboard out.
The question changes — “which shot wins more points?” becomes “which loans are underperforming?” or “where are we losing margin?” — but the build is the same. Ingest the source data. Compute the metrics that matter. Put them in front of the person who makes decisions.
If your business has data sitting in spreadsheets, scattered across tools, or locked in platforms — and you don't have a data team to make sense of it — this is what we build. One person with domain knowledge and AI, turning your data into tools you actually use.
Work with Nektar
Got data you've never been able to use? Let's turn it into something you actually look at. Free 30-minute data audit — no pitch, just a clear picture of what's possible.
Book a free data audit