Origin and detection of discrepancies between price and value
Research tools and the investment analysis they make possible, side by side. Everything here starts from primary sources and shows its work.
Focused applications for investment research: turning primary sources into things you can actually think with. All built in Rust and Python, with an emphasis on transparency and provenance over black-box outputs. These are evolving research tools: some are more mature than others, and all continue to be refined and expanded as new ideas and use cases emerge.
An auditable financial data layer built from SEC filings. It reconstructs the numbers behind 10-Ks and 10-Qs from XBRL and surfaces where the structured data and the HTML narrative disagree — so every figure is traceable back to its source document.
provenance.discrepancies.eu →The language of SEC 10-K filings across the S&P 500 — sentiment trends, Item 1A risk-factor evolution, keyword-in-context tone, full-text search, and a composable query workbench over a decade of narrative disclosure.
edgar-explorer.discrepancies.eu →Earnings-call analysis: tone shifts across quarters, hedging and uncertainty language, and the gap between the prepared remarks and how management actually answers questions in the Q&A.
subtext.discrepancies.eu →A cross-border ESEF filing explorer. It uses the shared IFRS taxonomy to line European companies up against one another across jurisdictions — same concepts, genuinely comparable figures.
meridian.discrepancies.eu →A monetary-policy transmission monitor. It pulls ECB and Bundesbank time series over SDMX, tracks data vintages, and follows how policy moves propagate — with every revision kept on the record.
leitwerk.discrepancies.eu →A European financial-data fragmentation mapper. It assembles a company profile from every available public source and makes the gaps, mismatches, and format inconsistencies visible rather than quietly papering over them.
mosaic.discrepancies.eu →The tools produce the insight; the writing puts it to work. Company analysis worked from primary sources, a portfolio tracked in the open, and the occasional longer piece on the framework itself.
Deep write-ups of individual public companies, worked from the filings and the transcripts: business, economics, valuation with explicit assumptions, specific risks, and a verdict — including the ones that conclude “good business, wrong price.” The apps appear at work throughout.
Read the case studies →A paper portfolio kept in the open. Rules published upfront and never changed retroactively; current holdings, closed positions with full return data, and performance against a stated benchmark. Accountability, not prediction.
See the record →Longer pieces on the framework itself — how computational analysis of filings changes fundamental research, what corpus analysis of Graham and Buffett reveals, and how US and European disclosure regimes compare on data quality.
Read the essays →I build financial-research infrastructure and explore value investing. One consistent interest is provenance: where financial data comes from, where it breaks, and how to make it transparent and independently verifiable. It’s the same question whether the output is an ingestion pipeline that reconciles a filing against its own XBRL, or a valuation that states every assumption it rests on. Another interest is the origins of value investing, as well as its latest techniques.
The habits come from a background in the humanities (philosophy and history), which encourages close reading and working from primary sources. I take these skills and direct them, along with my technical experience, at filings and market data. The tools here are built in Rust and Python; the writing follows analytical discipline. A small, self-directed body of work.
Full CV available on request. The code is on GitHub, or reach me by email.
Financial research is only as good as the data underneath it, and that data deserves to be transparent, traceable, and independently verifiable. The tools embody that principle technically, the case studies analytically. Insight the tools surface becomes writing; the friction found while writing feeds back into the tools.