Projects

Quantitative work you can explore.

Ordered by relevance, not date. Case studies explain work I have delivered and state my role; interactive demos show how I approach a problem and identify their data.

Featured quantitative work

Portfolio analytics, market and economic research, and supervisory risk work. Case studies state my role; demos state their data.

Investment analytics

Wealth analytics and macro research for an RIA

The analytics behind an RIA's investment process: monthly performance attribution, model-portfolio production, automated client reporting, and the macro-data pipeline and summary statistics behind a custom diffusion-index risk indicator that guides how the firm trades its managed accounts.

My role: Consultant via Second Difference Solutions

Diffusion-index risk indicatorPortfolio attributionTrade rebalancingModel portfoliosMacro data pipelineRPythonRedtail CRM

Read the case study

Economic and market data

Economic and market dashboards

I built R and Quarto dashboards around shared data pullers, cleaning functions, and charting helpers. The current site includes interactive examples with dated FRED snapshots.

My role: Author

RtidyversetidyquantFREDQuartoPlotly

Read the case study

Quantitative research

Signal research: a 147-rule sweep against buy and hold

I swept 147 technical trading rules in 6 indicator families on NVIDIA, choosing on 2000 to 2014 and testing on 2015 to 2026. 0 of 147 beat buy and hold on return, and the best Sharpe edge was smaller than the same search finds in noise.

My role: Author

PythonParameter sweepOut-of-sample testingPermutation testsWalk-forward

Read the case study

Interactive demo / Public data snapshot

Yield curve recession signal

The 10-year minus 2-year Treasury spread from FRED, back to 1976, with NBER recessions shaded. A dated public-data snapshot with interactive time ranges.

Try this: Compare Full history with Last 10 years. Check the observation date before interpreting the latest spread.

EconomicsFREDNext.js

Open the demo Read the write-up

Interactive demo / Synthetic data

Supervision and trade surveillance

Surveillance alert triage for a synthetic broker-dealer: the disposition funnel from alert to finding, open alerts aging against SLA, and a tuning panel that shows what raising a rule threshold really costs.

Try this: Compare West with All regions. Then raise the threshold and watch alert volume and missed findings change.

ComplianceRegTechNext.js

Open the demo Read the write-up

Interactive demo / Synthetic data

Portfolio metrics engine

Risk and return for a synthetic two-asset portfolio, recomputed in the browser by a dependency-free TypeScript engine: Sharpe, Sortino, drawdowns, VaR, beta, alpha, and capture ratios, with any window the history cannot cover shown as unavailable.

Try this: Choose the 10 year window, then move the weight slider to 100% Asset A and watch Unavailable turn into numbers.

Portfolio analyticsTypeScriptNext.js

Open the demo

Machine learning

A worked example on synthetic data, with the method visible.

Interactive demo / Synthetic data

Sales segmentation clustering

An interactive k-means demo on a synthetic distributor book. Toggle a revenue view to a cluster view and watch the high-margin accounts a revenue ranking hides light up.

Try this: Switch between Cluster view and Revenue view. Compare high-revenue accounts with high-margin segments.

ClusteringAnalyticsNext.js

Open the demo Read the write-up

Software and automation

Financial software and platforms I built or co-founded, and an operations dashboard on synthetic data.

SaaS product

BizAnalyzer

A QuickBooks-connected financial analytics platform: standardized statements, ratios, forecasts, and valuations, with multi-model AI commentary. I build and run it end to end through Second Difference Solutions.

My role: Founder and builder

Next.jsFirebasePythonMulti-LLMQuickBooksStripe

Read the case study 2diff.io

SaaS platform (joint venture)

Melody Map

A music-educator directory I co-founded and built end to end on Firebase: location search, subscriptions, and a pay-per-referral lead engine.

My role: Co-founder and technical lead

Next.jsFirebaseGoogle PlacesStripePython

Read the case study melodymap.io

Interactive demo / Synthetic data

Finance operations dashboard

Working capital, AR and AP aging, three-way match exceptions, and close status for a synthetic distributor absorbing an acquisition. Filter to the acquired branch and watch it converge.

Try this: Select Acquired branch. Compare its cash cycle, overdue receivables, and close status with All branches.

FinanceAutomationNext.js

Open the demo Read the write-up

Planned

Work I intend to do next. None of it exists yet; each card lists what it will cover.

Planned

Factor screener

A screener that will rank a stock universe on standard equity factors, so a reader can see which factors lead, which lag, and how the rankings shift over time.

Will cover

MomentumValueQualityProfitabilityGrowthVolatilitySize

Planned

Macro regime model

A model that will classify the economy as expanding, slowing, contracting, or recovering from public indicators, with each regime call dated and its inputs shown beside it.

Will cover

Yield curveLaborInflationProductionHousingPolicy ratesCredit

Planned

SEC filings data extension

An extension of the FRED pipeline behind the Data Lab that will add company data from SEC filings, with the same validation and dated snapshots the economic series use.

Will cover

SEC filingsCompany fundamentalsValidation checksDated snapshots

Enterprise data and BI labs

SQL, Databricks, and Power BI, worked hands-on

Each lab runs on a synthetic company and is written up as the deliverable: one margin bridge tied out across Databricks, SQL, and Power BI, and the change-control checklist for the report. They demonstrate the patterns, not production tenure with the tools.

One margin bridge, three tools

One monthly gross-margin bridge built in Databricks, traced with SQL, and recomputed in Power BI DAX, with every number tied out across all three tools.

Read the lab4 min read

A Power BI report under change control

A checklist for running a Power BI report like a control: .pbip source files, required review, two CI gates, a scripted publish, and an audit trail.

Read the lab2 min read

Hiring for investment, portfolio, or risk analytics?

Send me the role and the problem your team needs to solve. I can walk through the relevant work.

Discuss a role Read my resume

Featured quantitative work

Portfolio analytics, market and economic research, and supervisory risk work. Case studies state my role; demos state their data.

Investment analytics

Wealth analytics and macro research for an RIA

My role: Consultant via Second Difference Solutions

Diffusion-index risk indicatorPortfolio attributionTrade rebalancingModel portfoliosMacro data pipelineRPythonRedtail CRM

Read the case study

Economic and market data

Economic and market dashboards

I built R and Quarto dashboards around shared data pullers, cleaning functions, and charting helpers. The current site includes interactive examples with dated FRED snapshots.

My role: Author

RtidyversetidyquantFREDQuartoPlotly

Read the case study

Quantitative research

Signal research: a 147-rule sweep against buy and hold

My role: Author

PythonParameter sweepOut-of-sample testingPermutation testsWalk-forward

Read the case study

Interactive demo / Public data snapshot

Yield curve recession signal

The 10-year minus 2-year Treasury spread from FRED, back to 1976, with NBER recessions shaded. A dated public-data snapshot with interactive time ranges.

Try this: Compare Full history with Last 10 years. Check the observation date before interpreting the latest spread.

EconomicsFREDNext.js

Open the demo Read the write-up

Interactive demo / Synthetic data

Supervision and trade surveillance

Try this: Compare West with All regions. Then raise the threshold and watch alert volume and missed findings change.

ComplianceRegTechNext.js

Open the demo Read the write-up

Interactive demo / Synthetic data

Portfolio metrics engine

Try this: Choose the 10 year window, then move the weight slider to 100% Asset A and watch Unavailable turn into numbers.

Portfolio analyticsTypeScriptNext.js

Open the demo

Machine learning

A worked example on synthetic data, with the method visible.

Interactive demo / Synthetic data

Sales segmentation clustering

An interactive k-means demo on a synthetic distributor book. Toggle a revenue view to a cluster view and watch the high-margin accounts a revenue ranking hides light up.

Try this: Switch between Cluster view and Revenue view. Compare high-revenue accounts with high-margin segments.

ClusteringAnalyticsNext.js

Open the demo Read the write-up

Software and automation

Financial software and platforms I built or co-founded, and an operations dashboard on synthetic data.

SaaS product

BizAnalyzer

My role: Founder and builder

Next.jsFirebasePythonMulti-LLMQuickBooksStripe

Read the case study 2diff.io

SaaS platform (joint venture)

Melody Map

A music-educator directory I co-founded and built end to end on Firebase: location search, subscriptions, and a pay-per-referral lead engine.

My role: Co-founder and technical lead

Next.jsFirebaseGoogle PlacesStripePython

Read the case study melodymap.io

Interactive demo / Synthetic data

Finance operations dashboard

Working capital, AR and AP aging, three-way match exceptions, and close status for a synthetic distributor absorbing an acquisition. Filter to the acquired branch and watch it converge.

Try this: Select Acquired branch. Compare its cash cycle, overdue receivables, and close status with All branches.

FinanceAutomationNext.js

Open the demo Read the write-up

Planned

Work I intend to do next. None of it exists yet; each card lists what it will cover.

Planned

Factor screener

A screener that will rank a stock universe on standard equity factors, so a reader can see which factors lead, which lag, and how the rankings shift over time.

Will cover

MomentumValueQualityProfitabilityGrowthVolatilitySize

Planned

Macro regime model

A model that will classify the economy as expanding, slowing, contracting, or recovering from public indicators, with each regime call dated and its inputs shown beside it.

Will cover

Yield curveLaborInflationProductionHousingPolicy ratesCredit

Planned

SEC filings data extension

An extension of the FRED pipeline behind the Data Lab that will add company data from SEC filings, with the same validation and dated snapshots the economic series use.

Will cover

SEC filingsCompany fundamentalsValidation checksDated snapshots