Prism Data Lab https://prismdatalab.com/ Financial data analysis with the code and dataset attached, and a clear-eyed look at where language models fit in finance: research, filings, risk, and limits. 2026-10-10T00:00:00Z Month-of-year effects in market data, under a false discovery rate https://prismdatalab.com/posts/month-of-year-effects-false-discovery-rate.html https://prismdatalab.com/posts/month-of-year-effects-false-discovery-rate.html Sixty month-of-year tests on the Nasdaq, Treasuries, oil, gasoline and natural gas: 10 pass at p < 0.05, and a 5% false discovery rate keeps 4, all gasoline. 2026-10-09T00:00:00Z Real versus nominal: deflating a series correctly, and six mistakes https://prismdatalab.com/posts/real-versus-nominal-deflating-correctly.html https://prismdatalab.com/posts/real-versus-nominal-deflating-correctly.html From January 2021 to August 2026, real hourly pay fell 0.8% by the CPI and rose 1.7% by the PCE price index. How to deflate, checked against the BLS, and six traps. 2026-10-06T00:00:00Z The CPI basket explained, from the BLS relative importance tables https://prismdatalab.com/posts/cpi-basket-relative-importance.html https://prismdatalab.com/posts/cpi-basket-relative-importance.html Shelter is 35.6% of the CPI-U basket for 2026 and energy 6.4%, yet energy supplied 38% of the index's rise from December to August. The weights, and how to use them. 2026-09-25T00:00:00Z An event study around 199 FOMC announcements, honestly scored https://prismdatalab.com/posts/fomc-event-study-announcement-window.html https://prismdatalab.com/posts/fomc-event-study-announcement-window.html Rates, VIX and the Nasdaq in a five-day window around every FOMC statement since 2003, with permutation tests, a multiple-testing correction, and what survives it. 2026-09-22T00:00:00Z Correlation instability: rates, equities, and the dollar since 2006 https://prismdatalab.com/posts/correlation-instability-rates-equities-dollar.html https://prismdatalab.com/posts/correlation-instability-rates-equities-dollar.html Rolling 126-day correlations between the 10-year yield, equities, and the dollar, with a simulation that separates a real regime change from sampling noise. 2026-09-18T00:00:00Z Data revisions in FRED: ALFRED vintages and why final is not final https://prismdatalab.com/posts/alfred-vintages-data-revisions.html https://prismdatalab.com/posts/alfred-vintages-data-revisions.html Every payroll and real GDP vintage since 2010, pulled from ALFRED without a key: how far first prints moved, the annual benchmarks, and the level trap. 2026-09-15T00:00:00Z Bond math from first principles: price, yield, duration, convexity https://prismdatalab.com/posts/bond-math-price-yield-duration-convexity.html https://prismdatalab.com/posts/bond-math-price-yield-duration-convexity.html Price, yield to maturity, Macaulay and modified duration, and convexity derived and coded from scratch, then applied to the Treasury par curve of 2026-09-09. 2026-09-11T00:00:00Z Value at Risk three ways, backtested on the S&P 500 https://prismdatalab.com/posts/value-at-risk-three-ways.html https://prismdatalab.com/posts/value-at-risk-three-ways.html Historical, parametric normal, and Cornish-Fisher VaR computed on ten years of FRED's S&P 500 series, then backtested day by day with the Kupiec coverage test. 2026-09-08T00:00:00Z A 2026 10-K's AI disclosure, quoted and dated: Target Corporation https://prismdatalab.com/posts/target-10-k-2026-ai-disclosure.html https://prismdatalab.com/posts/target-10-k-2026-ai-disclosure.html Every sentence mentioning AI in Target's 10-K filed 2026-03-11 (accession 0000027419-26-000016), counted by section and quoted, with a script for any filing. 2026-09-06T00:00:00Z Reading SEC EDGAR in Python: submissions, full-text search, XBRL facts https://prismdatalab.com/posts/sec-edgar-python-submissions-fts-xbrl.html https://prismdatalab.com/posts/sec-edgar-python-submissions-fts-xbrl.html The three free EDGAR JSON endpoints in one standard-library script: a company's filings by accession number, full-text search hits, and deduplicated XBRL facts. 2026-09-06T00:00:00Z Log returns, simple returns, and the compounding errors people make https://prismdatalab.com/posts/log-returns-vs-simple-returns.html https://prismdatalab.com/posts/log-returns-vs-simple-returns.html Simple and log returns on ten years of FRED's S&P 500 series: the sum identity, variance drag, the size-of-move gap, and the mistakes each choice invites. 2026-09-06T00:00:00Z Language models reading financial statements: published accuracy https://prismdatalab.com/posts/llms-reading-financial-statements-accuracy.html https://prismdatalab.com/posts/llms-reading-financial-statements-accuracy.html Four 2026 benchmarks measured LLMs computing ratios, verifying statements, answering from full 10-Ks, and summarising MD&A. Accuracies and failure modes, cited. 2026-09-06T00:00:00Z Time-series forecasting with LLMs: claims, benchmarks, and sceptics https://prismdatalab.com/posts/llm-time-series-forecasting-claims.html https://prismdatalab.com/posts/llm-time-series-forecasting-claims.html What 2026 papers report when language and time-series foundation models forecast returns, volatility, and factor rankings against simple baselines, by arXiv id. 2026-09-06T00:00:00Z Extracting tables from filings: a model versus XBRL, when each wins https://prismdatalab.com/posts/llm-table-extraction-vs-xbrl.html https://prismdatalab.com/posts/llm-table-extraction-vs-xbrl.html XBRL company facts for one filer, deduplicated, with the tag changes and restatements any extractor hits, and what 2026 table-extraction benchmarks report. 2026-09-06T00:00:00Z Where free market data actually comes from, and what each source lags https://prismdatalab.com/posts/free-market-data-sources-and-lags.html https://prismdatalab.com/posts/free-market-data-sources-and-lags.html Eight FRED series traced to the agency or exchange that produces them, with native frequency and the days between last observation and retrieval date, computed. 2026-09-06T00:00:00Z Pulling FRED series without an API key, cleanly, in Python https://prismdatalab.com/posts/fred-series-python-no-api-key.html https://prismdatalab.com/posts/fred-series-python-no-api-key.html FRED serves every series as a plain CSV with no key. A standard-library script that downloads, merges, and coverage-checks any list of series ids, with the gotchas. 2026-09-06T00:00:00Z The August 2026 jobs report as it appears in FRED, in five paragraphs https://prismdatalab.com/posts/fred-release-brief-2026-09-05.html https://prismdatalab.com/posts/fred-release-brief-2026-09-05.html A dated reading of the 2026-09-04 Employment Situation through four FRED series retrieved the next day: payrolls, unemployment, participation, and earnings. 2026-09-06T00:00:00Z Earnings-call sentiment: what the papers measured, FinBERT to LLMs https://prismdatalab.com/posts/earnings-call-sentiment-finbert-to-llms.html https://prismdatalab.com/posts/earnings-call-sentiment-finbert-to-llms.html What financial-sentiment research reports, from the 2019 FinBERT paper to 2026 studies of call tone, evasion, KPI extraction, and lookahead bias, by arXiv id. 2026-09-06T00:00:00Z BLS data in Python: CPI components and payrolls without a key https://prismdatalab.com/posts/bls-cpi-components-python-no-key.html https://prismdatalab.com/posts/bls-cpi-components-python-no-key.html The BLS public API v1 needs no key. A script pulls CPI components and the jobs headline, handles the "-" returned for a missing month, and prints 12-month changes. 2026-09-06T00:00:00Z What companies say about AI in 10-K risk factors: an EDGAR read https://prismdatalab.com/posts/ai-in-10-k-risk-factors-edgar-full-text.html https://prismdatalab.com/posts/ai-in-10-k-risk-factors-edgar-full-text.html 10-K documents matching four AI phrases per year since 2019 from EDGAR full-text search, a 40-filing sample with accession numbers, and what the counts cannot say. 2026-09-06T00:00:00Z The 10-year minus 2-year spread since 1976: every inversion, dated https://prismdatalab.com/posts/yield-curve-inversions-since-1976.html https://prismdatalab.com/posts/yield-curve-inversions-since-1976.html Every inversion of the 10-year minus 2-year Treasury spread since the 2-year series began in 1976, with dates, depth, duration, and what followed, from FRED. 2026-09-05T00:00:00Z VIX regimes: how often the index closes above 30, by year since 1990 https://prismdatalab.com/posts/vix-closes-above-30-by-year.html https://prismdatalab.com/posts/vix-closes-above-30-by-year.html Every calendar year since 1990 with the number of VIX closes above 20, 30, and 40, the yearly maximum and its date, and the longest stretches above 30, from FRED. 2026-09-05T00:00:00Z Sharpe, Sortino, and max drawdown computed by hand on the S&P 500 https://prismdatalab.com/posts/sharpe-sortino-max-drawdown-by-hand.html https://prismdatalab.com/posts/sharpe-sortino-max-drawdown-by-hand.html Sharpe, Sortino, and maximum drawdown worked step by step in Python on ten years of FRED's S&P 500 series, with a 3-month Treasury risk-free rate, year by year. 2026-09-05T00:00:00Z Sahm rule from FRED data: unemployment and recessions since 1960 https://prismdatalab.com/posts/sahm-rule-unemployment-recessions.html https://prismdatalab.com/posts/sahm-rule-unemployment-recessions.html The Sahm rule recomputed from the unemployment rate in Python and checked against the real-time series and NBER dates, including the 2024 trigger with no recession. 2026-09-05T00:00:00Z Mortgage rates and the 10-year Treasury: the spread since 1971 https://prismdatalab.com/posts/mortgage-rate-10-year-spread.html https://prismdatalab.com/posts/mortgage-rate-10-year-spread.html The 30-year fixed mortgage rate minus the 10-year Treasury yield, every survey week since 1971, by decade and year, with the widest and narrowest weeks. 2026-09-05T00:00:00Z M2 and inflation since 1960: what the data shows and what it does not https://prismdatalab.com/posts/m2-and-inflation.html https://prismdatalab.com/posts/m2-and-inflation.html Twelve-month M2 growth against CPI inflation from 1960 to July 2026: lagged correlations by period, the 2021 peak, the 2023 contraction, and the limits. 2026-09-05T00:00:00Z Fed funds versus CPI since 1955: the real rate, decade by decade https://prismdatalab.com/posts/fed-funds-vs-cpi-real-rates.html https://prismdatalab.com/posts/fed-funds-vs-cpi-real-rates.html The effective federal funds rate minus CPI inflation, monthly from 1955 to July 2026, summarised by decade with the longest negative stretches, from FRED. 2026-09-05T00:00:00Z Backtest overfitting: the deflated Sharpe ratio and a simulation https://prismdatalab.com/posts/backtest-overfitting-deflated-sharpe.html https://prismdatalab.com/posts/backtest-overfitting-deflated-sharpe.html Why the best of many backtests looks good even when every strategy is noise: a simulation you can run, plus the deflated Sharpe ratio applied to it. 2026-09-05T00:00:00Z