Market research meets data storytelling. Tracking brands, testing hypotheses, and finding patterns others miss—using AI.
From a data nerd: experiments and tracking dashboards
Lessons from analyzing millions of rows with AI tools
Don't upload data and say 'analyze this.' Ask a specific question: 'What patterns exist in my spending over time?' Focused questions get useful answers.
Clean CSVs with headers are AI's best friend. Avoid PDFs, images of tables, or nested JSON. The cleaner your input, the better your output.
Always ask: 'How did you calculate that?' or 'What assumptions did you make?' AI will confidently give you wrong answers if you don't verify the approach.
If AI finds something unexpected, that's either a breakthrough or a mistake. Ask it to double-check, or run the same question a different way.
First pass is rarely perfect. Say 'That's interesting, but can you break it down by month?' or 'Now compare that to X.' Build toward insight.
Watch for hedging language: 'might,' 'could suggest,' 'possibly.' That's AI telling you it's uncertain. Push for confidence levels or caveats.
15 years in market research taught me that the best insights come from connecting signals others ignore. These projects blend public data with editorial perspective—not to replace expensive tools, but to show what's possible when you ask better questions. More about me →