Skip to content

Quant Research & Alt Data

Satellite imagery signals, credit card transaction data, web-scraped sentiment, supply chain networks, Google Trends, social bot filtering, transcript NLP.

Skills
20
Hand off
121
Handed off from
76

Index

All 20 skills in this domain, alphabetically. Hover a row for its situation; open it for the full playbook.

A

alternative-data-vendor-due-diligence-checklistUse when onboarding an alternative data vendor, triaging legal rights, material non-public information exposure, scraping and computer-misuse risk, terms of service and personal data handling before the dataset reaches research.alternative-dataapp-download-and-usage-data-for-consumer-companiesUse when turning mobile app engagement panels such as downloads and active users into a fundamental signal for consumer companies, with the panel normalisation and lag that makes the estimate point-in-time defensible.dau-mau

B

backtesting-alt-data-strategies-with-realistic-availability-lagUse when backtesting an alternative data strategy, to key every observation on its publication date rather than its event date, because the gap between the two is where impossible backtest results come from.lookahead-bias

C

central-bank-communication-nlp-analysisUse when quantifying the hawkish or dovish stance of central bank statements, minutes and press conferences for a macro, rates or FX strategy, with sentence-level negation handling and policy uncertainty scoring.sentiment-analysiscredit-card-transaction-data-signal-constructionUse when forecasting a consumer company's quarterly revenue from card transaction panels ahead of the earnings release, normalising panel bias and computing year-on-year growth against consensus.credit-card-data

E

earnings-call-transcript-nlp-signal-researchUse when building an equity feature from earnings call transcripts, scoring financial sentiment with a finance-specific lexicon and measuring the tone gap between scripted remarks and the spontaneous analyst question session.earnings-transcriptsesg-data-signal-research-and-vendor-comparisonUse when ESG ratings for the same issuer disagree across MSCI, Sustainalytics and LSEG, producing a consensus score plus an explicit dispersion measure of how much the vendors disagree.alternative-data

F

factor-research-multiple-testing-correctionUse when a research pipeline screens many candidate factors and selection rests on t-statistics or p-values. Bonferroni and Holm bound family-wise error; Benjamini-Hochberg and BHY bound false discovery rate under dependence; plus the Harvey-Liu-Zhu t-hurdle.multiple-testing-correction

G

google-trends-and-search-volume-signal-researchUse when building a retail-attention feature from search volume, standardising against a strictly trailing baseline and enforcing the publication lag with an explicit as-of date. A research feature, not a sized signal.svi

I

insider-transaction-filing-signal-researchUse when building an equity factor from SEC Form 4 insider filings, scoring open-market purchases and sales on the EDGAR dissemination timestamp rather than the trade date and weighting by the filer's role.sec-form-4

J

job-posting-data-as-a-growth-signalUse when building a fundamental growth feature from web-scraped job postings, measuring quarter-on-quarter hiring velocity and engineering role mix while penalising stale ghost listings.alt-data

O

options-flow-unusual-activity-detectionUse when screening an options print feed for trades large relative to the series open interest and average volume carrying six-figure premium, and inferring buy or sell side from the quote. Every label is an inference.unusual-activity

P

patent-filing-data-for-innovation-signal-researchUse when building a cross-sectional innovation factor from patent filings, scoring each from the date it actually became public and requiring per-date citation snapshots so the factor stays point-in-time.uspto

R

research-environment-vs-production-environment-parityUse when promoting a model or feature pipeline out of notebooks into live execution and the question is whether production computes the same number, auditing runtime, package and data-path parity vectors. Deployment gating is environment-parity-dev-staging-production.research-vs-productionresearch-idea-pipeline-tracking-and-prioritizationUse when a research backlog is longer than the capacity to work it, registering each hypothesis once and ranking the active queue by a transparent triage score so ordering does not follow whoever argued most recently.research-pipeline

S

satellite-imagery-based-signal-researchUse when a processed Earth-observation metric such as a parking-lot vehicle count, tank fill fraction or vegetation index must become a z-scored research signal with an explicit availability lag. Not a computer-vision pipeline.alternative-datasocial-media-sentiment-signal-with-bot-filteringUse when building a signal from social media posts and the stream must be screened for bots, spam and coordinated campaigns before aggregation. Metadata hygiene rules, not a bot detector.bot-filteringsupply-chain-data-for-earnings-predictionUse when estimating a company's quarterly revenue growth from the reported results of firms one step up and down its supply chain, reading supplier growth through concentration weights and customer inventory as a drag.alternative-data

W

weather-data-signal-research-for-commodity-strategiesUse when researching weather signals for energy or agricultural futures, computing population-weighted heating and cooling degree days and growing degree days from station or forecast data. Pricing weather derivatives is a separate skill.weather-signalsweb-scraped-sentiment-data-pipelineUse when scraped financial text must become a point-in-time sentiment feature for a ticker, cleaning markup and entity noise, scoring against a verified finance lexicon and deduplicating syndicated copy.sentiment-analysis