I build the data systems and analytical tools that shorten the distance between raw information and a good decision.
Valuable information often lives across transcripts, documents, product catalogues, customer records, market data, APIs, and operational systems. Before it can answer useful questions, it usually needs to be gathered, normalized, connected, and made interpretable.
My work spans that full path: data collection, transformation, analysis, and lightweight tools for exploring the result. Language analysis is one area of particular depth, but the underlying problem is broader: creating reliable structure around information so people can use it well.
Collecting data from files, APIs, websites, and internal systems; normalizing it into durable datasets; and designing outputs that are straightforward to query, share, and build on.
Exploring pricing, customer, market, operational, or language data to surface patterns worth acting on - with an emphasis on interpretable results rather than analysis for its own sake.
Structured analysis of transcripts, reviews, customer communications, and other language data, alongside practical AI workflows and tools where they genuinely improve the work.
Language delta analysis revealing which import agency portfolios carried the most distinctly negative reception - before sales data reflected the problem.
A research-grade corpus built from raw JSON - Q&A sections extracted, executive voices separated, delivered in multiple formats for layered institutional analysis.
Resonant Analytics was founded in Montreal by Jack Geddes, a practitioner with deep experience in behavioral language analytics - building data pipelines, leading technical client engagements, and working at the intersection of language, data, and commercial outcomes.
That background shapes how I approach the work. I don’t just move data from one format to another - I think carefully about what structure will make the output most useful downstream, and what the output needs to enable for the people who will act on it.
If you're dealing with unstructured text at scale and not getting the intelligence you need from it, I'd like to hear about it.