Build a Fiscal Announcement-to-Spending Evidence Tracker

A public expenditure analyst monitoring program delivery.

Track a fictional program from announced budget through awards, obligations, and outlays. Explain why a headline funding amount differs from recorded implementation.

Key: award_id + transaction_id + modification_id.

Fields: obligation_change, outlay_amount, ceiling, action_date, fiscal_year, performance_geography.

Keep recipient address separate from place of performance. Do not sum repeated award snapshots.

Map awards to the defined program scope.

Reconcile obligations using signed transaction changes.

Compare implementation stages without mixing amounts.

Export a funding timeline with scope and data-quality exceptions.

A program timeline, award drilldown, geography view, and reconciliation queue.

A transaction-level spending reconciliation pipeline.

A report separating announced, obligated, and paid amounts.

Runnable local application, setup commands, and a dependency lockfile.

Versioned input data, data dictionary, acquisition instructions, and rights notes.

Calculation traces, source references, assumption records, and exported review artifacts.

A deterministic baseline and tests for normal, incomplete, contradictory, and failure-recovery cases.

A recorded complete workflow with a reproducible manifest and explicit limitations.

Deobligations reduce net obligations.

Repeated snapshots cannot inflate spending totals.

Do not infer local economic impact from recipient addresses.

Missing outlays remain unavailable, not equal to obligations.

A later modification reduces an award. The net amount falls even though the original announcement remains unchanged.

The attached starter files are prototype inputs. Expected starter answers are visible. They are not independent holdouts.

Full target: Acquire a bounded public award sample with transaction modifications. Add a fictional program announcement and clearly labeled outlay records if unavailable at the same grain.

Complete target contracts and expert review remain pending. Public observations are not historical release vintages.

DuckDB, pandas, Streamlit

Use an available model appropriate to the task. Record provider, model identifier, version or retrieval date, settings, cost, and latency. Use deterministic code for calculations. Models may extract, retrieve, or explain evidence. They must not invent missing records, override approval rules, or claim professional authority.

Administrative spending records do not establish program effectiveness or fraud.

No fabricated records, source citations, professional endorsements, or expert-adjudication claims.

No future-data leakage into dated decisions, model fitting, or historical valuation inputs.

No unauthorized payments, account restrictions, regulatory filings, or production changes.

No automated appraisal certification, tax-compliance determination, or unsupported fraud accusation.

No concealed uncertainty or silently invented inputs required for a numerical conclusion.

Correct domain calculations and justified methods with documented assumptions. Weight: 30.

Source lineage, temporal integrity, rights checks, and explicit synthetic-data labels. Weight: 25.

A working application with review, export, error handling, and reproducible state. Weight: 20.

Adversarial cases, uncertainty, missing-data behavior, and sensitivity analysis. Weight: 15.

Evidence-linked conclusions that distinguish facts, assumptions, and unresolved judgments. Weight: 10. Score each dimension from zero to two. Weighted score = sum(weight * score / 2). A failed safety gate blocks acceptance.