DataSynth 3.0

Counterfactual Simulation for Financial Data

Run what-if scenarios on structurally coherent synthetic data. Intervene on any variable in a 23-node causal DAG and observe downstream effects across the financial process chain.

11 Scenario Packs

5 fraud, 2 control failure, 3 macro shock, and 1 operational scenario — each with configurable severity

Invoice Fraud Escalation

Fraud

Duplicate invoices, shell vendors, amount manipulation with progressive detection evasion

Payroll Ghost Employees

Fraud

Phantom employees, salary siphoning, and terminated-but-active payroll entries

Expense Reimbursement Abuse

Fraud

Split transactions, fictitious receipts, and policy threshold gaming

Revenue Recognition Manipulation

Fraud

Channel stuffing, bill-and-hold, and premature cutoff scenarios

Related-Party Concealment

Fraud

Hidden relationships, circular transactions, and beneficial ownership obfuscation

Segregation of Duties Failure

Control

SoD violations across approval chains with compensating control gaps

Access Control Breakdown

Control

Privilege escalation, orphaned accounts, and emergency access abuse

Interest Rate Shock

Macro

Parallel shift, twist, and inversion scenarios propagated through asset valuations and cash flows

FX Volatility Spike

Macro

Currency pair dislocations with hedging ineffectiveness and translation exposure

Credit Spread Widening

Macro

Sector-specific spread blowouts with contagion to counterparty credit risk

System Migration Failure

Operational

Data mapping errors, cutover timing mismatches, and reconciliation breaks during ERP migration

23-Node Financial Process DAG

Every intervention propagates through a directed acyclic graph with 25 causal edges across 6 categories

Revenue Chain

4 nodes

Sales orders, invoicing, revenue recognition, cash collection

Procurement Chain

4 nodes

Purchase requisition, PO approval, goods receipt, invoice matching

Treasury Chain

4 nodes

Cash position, FX exposure, hedging, settlement

Credit Risk Chain

4 nodes

Counterparty rating, exposure calculation, provision, write-off

Control Chain

4 nodes

Authorization, segregation, reconciliation, reporting

Macro Environment

3 nodes

Interest rates, FX rates, credit spreads

Python SDK Example

import vynfi

client = vynfi.Client(api_key="vf_live_...")

# Define a counterfactual scenario
job = client.simulate(
    scenario_pack="interest_rate_shock",
    base_dataset="manufacturing_ledger",
    interventions={
        "interest_rate": {"shift": "+200bp", "curve": "parallel"},
        "credit_spread": {"sector": "industrials", "widen": "+150bp"},
    },
    rows=50_000,
    propagate=True,  # Follow causal DAG
)

# Compare base vs counterfactual
diff = job.compare(metrics=["revenue", "provisions", "cash_flow"])
print(f"Revenue impact: {diff['revenue']['pct_change']:.1f}%")
print(f"Provision increase: {diff['provisions']['pct_change']:.1f}%")

Stop guessing — start simulating

Run counterfactual scenarios on causal financial data. 5,000 free credits to start.

View Pricing