Chicago's Fraud Detection Initiative

Cluster of high‑count days points to a period‑wide processing issue.

Spend rose from $500 to $4.9 M year‑over‑year, a >9,700% increase.

Year‑over‑year spend grew from $50,053 to $128.8 M, indicating a massive budget deviation.

Year‑to‑year analysis shows a different highest‑spend vendor, suggesting a procurement change.

Top spend vendor differed from prior year, indicating possible contract award shift.

Aggregating payments shows the five largest vendors together account for over two‑fifths of total disbursements, indicating high concentration risk.

Service‑level aggregation shows Public Improvement far exceeds all other categories, suggesting potential misallocation toward capital projects.

High frequency and large average payouts raise red flags for possible overpayment or lack of competitive bidding.

Cross‑dataset vendor overlap suggests possible double‑dipping or shared contracts across programs.

Date anomalies indicate data quality issues that could affect timing analyses and audit trails.

Payments linked to debarred entities show significant spend despite ineligibility.

Contract awards to ineligible firms indicate procurement compliance breaches.

Repeated payments over multiple years despite criminal fraud debarment.

Zero or negative award amounts for a debarred roofing firm raise flag for data integrity or unauthorized contracts.

Payment occurred before formal debarment, indicating possible retroactive enforcement gaps.

Low penetration suggests most debarred firms are effectively excluded, but high spend on few indicates risk concentration.

Slightly higher contract exposure than payments, indicating possible pre‑award vetting failures.

Repeated identical amounts across many dates suggest systematic duplicate billing.

Same voucher number appears twice with identical amount, a classic duplicate payment sign.

Extremely high duplicate count indicates possible systematic overpayment or data error.

Multiple unrelated vendors sharing a single address suggests a possible shell‑company hub.

Zero‑value contracts are atypical and may be placeholders for undisclosed payments.

High frequency of same‑amount payments suggests batch processing that may conceal irregularities.

Outlier detection identified a tiny set of salaries far above typical compensation, indicating possible unauthorized payments.

Extreme variance within a single role suggests irregular adjustments or collusion.

Highest compensated positions are singular executive roles, which may be justified but should be reviewed for policy compliance.

Department-level pay is broadly consistent across the city, limiting systemic bias.

While the absolute number is small, the presence of any outliers signals a risk that should be monitored.

Overpay distinct vouchers 71,007 vs matched distinct 111,748 shows a majority of payments are higher than the authorized award.

194,626 payment rows have no contract match, indicating possible data gaps or undocumented spend.

151,229 rows where payment vendor differs from contract vendor suggest possible invoicing errors or fraud.

69,241 rows where payment department differs from contract department could indicate scope violations.

26,681 contracts with $0 award may be data entry errors or missing approvals, risking uncontrolled spend.

High volume of orphan payments reduces auditability and may hide unauthorized spend.

License term expiration dates are all future, so no inactive licenses to flag.

A sizable number of active contracts are tied to soon‑to‑expire licenses, posing compliance risk.

Payments continue to flow to entities whose legal authority will lapse soon, increasing fraud exposure.

One vendor accounts for over 60% of near‑expiration contracts, a concentration risk.

Broad vendor base suggests many low‑volume relationships; focus can be narrowed to those with expiring licenses.

A massive count of distinct LICENSE IDs at one address signals multiple entities using the same location, a classic sign of address spoofing or shell companies.

High vendor density on a single street raises suspicion of shared facilities or shell entities rotating addresses to mask true ownership.

An unusually high number of distinct purchase orders for a single vendor may indicate a conduit used to channel funds through a preferred supplier.

A single vendor securing thousands of contracts can act as a hub for multiple shell entities, inflating spend and obscuring true beneficiaries.

Matching addresses across independent datasets ties a municipal license holder to a procurement vendor, revealing a possible undisclosed relationship.

Exact name joins between payment ledger and debarment list returned zero matches, indicating no direct hits.

Multiple change orders with a zero baseline and a massive final award suggest unchecked scope expansion.

A tiny initial award ballooned dramatically, flagging potential inflated change orders.

Zero‑baseline contracts that later receive multi‑million awards reveal possible bypass of competitive bidding.

Negative minima suggest adjustments that may mask prior overpayments or irregularities.

Frequent revisions correlate with larger spend, suggesting unchecked scope creep.

Across 450,594 vouchers, 171,836 have amounts with no cents, indicating a high prevalence of round figures.

The department processed 5,473 vouchers, of which 94.7% are round amounts, far above the overall average.

Among 1,147 vouchers from this vendor, 657 are round amounts, indicating a strong rounding pattern.

Across the ledger, 17,870 payments are exact multiples of 1,000, a relatively low occurrence.

Departments with only one voucher (e.g., LIC COMM & LOCAL LIQ CTRL COMM) have 100% round payments, suggesting ad‑hoc entries.

A large share of tiny payments aggregates to a sizable total, suggesting structuring.

High proportion of micro‑transactions contributes millions, a classic structuring sign.

Even with a massive overall spend, a quarter are tiny payments, indicating possible evasion of thresholds.

Significant share of micro‑payments creates a notable aggregate, flagging structuring risk.

Even a modest count of micro‑payments adds up within a huge overall spend, suggesting deliberate splitting.

Aggregating all award and payment amounts shows the five largest vendors together represent a modest share of overall expenditures.

All recorded spend for the Department of Zoning is tied to a single vendor, indicating no competition.

The entire budget for this board is paid to a single vendor, suggesting a potential single‑source arrangement.

All payments for this department flow to one vendor, indicating lack of competitive bidding.

A single vendor receives the entire spend of a major finance department, raising concerns about procurement controls.

Negative and >100% values suggest data errors or mis‑recorded refunds, which could conceal true vendor relationships.

Avg award exceeds twice the overall average ($817k) indicating unusually large payouts per transaction.

Average voucher amount far exceeds the overall average ($242k), indicating high‑value outliers.

City‑level aggregation shows Chicago accounts for the vast majority of award dollars.

Department‑level total reveals Aviation leads in absolute spending, raising exposure risk.

Top vendor spend suggests potential over‑reliance or pricing anomalies.

Overall emergency count is 1,764 of 182,314 contracts (0.188%). The Streets department shows 7.59% emergency rate across 2,504 contracts.

Only 66 sole‑source contracts (0.07% of total). Procurement Services department accounts for 1.107% of its 542 contracts.

Flagged spend $1.489 B versus total spend $149.008 B yields 0.037% flagged spend, indicating limited monetary exposure.

Top 10 flagged vendors each have 1‑2 total contracts, all flagged, suggesting risk from low‑volume, possibly opportunistic suppliers.

30% emergency rate (3 of 10 contracts) is an outlier but based on a very small pool, warranting targeted review.

A single tax identifier appearing under different names suggests possible ID reuse or misreporting, a red flag for fraud.

One business appearing under several tax IDs may indicate split reporting or shell entities, increasing audit risk.

When a licensed business address aligns with a procurement vendor but the legal name differs, it suggests deliberate obfuscation of the true vendor identity.

Multiple name variations for a single ID amplify suspicion of identity manipulation or data entry errors.

A single brand operating under multiple IDs may be structuring contracts to evade scrutiny.

Widespread name divergence across thousands of entities indicates systemic issues rather than isolated errors.

Absence of a corresponding business license for a vendor raises questions about legitimacy and compliance.

High variation density may pinpoint entities with aggressive identity splitting, warranting deeper audit.

Comparing each contract’s award to its department’s 95th‑percentile (p95) shows six outliers >2× p95, far above the overall average $0.82 M.

Their p95 values (e.g., $1.04 M for Public Health, $20.98 M for Finance) far exceed other departments, indicating systemic higher spend levels.

Large revision counts combined with max/min award disparity suggest scope creep or data anomalies.

These vendors consistently receive high‑value awards, warranting deeper vendor‑performance review.

Extended duration combined with high award values raises risk of scope expansion and budget strain.

A tiny fraction of contracts drives disproportionate fiscal exposure, highlighting concentration risk.

Count exceeds avg+3σ (347 ± 1014), indicating an extreme burst.

Three‑day sum exceeds avg+3σ, showing coordinated timing.

Average amount far exceeds overall avg $82k, flagging high‑value outliers.

Repeated occurrences indicate systemic risk of large‑value payments.