Skip to main content

Macro by Mark

HomeDashboardsSearchIndicatorsNewsCalendarDashboard BuilderMacroHelpAbout
Macro by Mark
Home
Indicators
Categories
GrowthPrices & InflationLabor MarketMonetary & Financial ConditionsNowcasting & Leading Indicators
Dashboards
News
Calendar
Search
Macro
Macro
OverviewHistoryConceptsModelsGlossary
About

Data-Driven Models

Loading Data-Driven Models

Macro by Mark

U.S. macro data with release timing, boards, and macro context.

Public U.S. data from agencies and market feeds.

MarkJayson.com↗Contact↗

Main

HomeIndicatorsDashboardsNewsCalendarSearch

Macro

MacroHistoryConceptsTheory-Based ModelsData-Driven ModelsModelsGlossary

About

HelpAboutMark Jayson Farol↗Privacy PolicyTerms of UseEthics & Compliance
LinkedInGitHubGoogle ScholarORCIDResearchGate

© 2026 Mark Jayson Martinez Farol

Models / Route

Data-first family

Data-Driven Models

This family now runs as one route. Enter through data, choose a model class, evaluate the setup, then move into forecast output.

Help
All ModelsData-Driven ModelsDataModelEvaluateForecast
Route notes

Connect one provider series for the live ARIMA path or stage an uploaded dataset for review.

VAR, BVAR, Ridge, LASSO, and Ensemble are fully modeled in the workflow but remain non-runnable in this pass.

ARIMA is the only live run path mapped onto the existing forecast engine.

Start
Data
Model
Configure
Evaluate
Forecast

Data-Driven Models

Build a Data-Driven forecast workflow

Connect a live macro series or stage an uploaded dataset, review the aligned frame, choose one model class, then move into evaluation and forecast output.

Entry paths

2

Live run class

ARIMA

Modeled classes

6

Upload a dataset

Stage a CSV first when the modeling frame lives outside the site registry.

Continue

Connect a provider series

Use the live registry path when you want a series that can move into the ARIMA run flow now.

Continue

Model classes

ARIMA

One-series baseline forecast built from historical persistence and differencing.

VAR

Interacting series with lag structure and stability checks.

BVAR

VAR structure with priors and shrinkage over a compact macro system.

RIDGE

Regularized regression for broad feature packs with stable shrinkage.

LASSO

Sparse regression that can zero out weaker predictors.

ENSEMBLE

Combines several empirical models into one scored baseline.