Intelligence Briefing
Key findings from across the data warehouse
--
Competitive races where the model gives the frontrunner <65% win probability
--
Power brokers donating to 10+ different filers — network hubs
--
Major donors ($100K+) active since 2020 — recent money in play
Consulting Practice Revenue
Client pipeline and delivery metrics
Monthly Recurring
--
--
Active Clients
--
--
Active Engagements
--
--
Deliveries This Month
--
--
Overdue
--
past scheduled cadence
Most Competitive Races Toss-ups
| Year | Office | Category | Candidates | Frontrunner Confidence |
|---|---|---|---|---|
| Loading competitive races... | ||||
Power Brokers Network
| Donor | Filers | Total $ | Community | Flags |
|---|---|---|---|---|
| Loading... | ||||
Recent Major Donors 2020+
| Donor | Total $ | Filers | City |
|---|---|---|---|
| Loading... | |||
Fundraising Benchmark Alerts Benchmarks
Under-Benchmarked Frontrunners
Strong win probability but below winner median fundraising
| Year | Office | Candidate | Win % | Gap |
|---|---|---|---|---|
| Loading... | ||||
Over-Performing Challengers
Non-incumbents fundraising above the 75th percentile
| Year | Office | Candidate | % of P50 | Tier |
|---|---|---|---|---|
| Loading... | ||||
Pacing Intelligence Trajectory
Competitive races behind pace and biggest pacing improvements
Behind-Pace Competitive Races
Frontrunners (win% > 35%) who are behind pace
| Year | Office | Candidate | Win % | Pctl | Flag |
|---|---|---|---|---|---|
| Loading... | |||||
Biggest Pace Improvements
Candidates who improved pacing percentile by 5+ points
| Year | Candidate | Prior | Current | +Δ | Flag |
|---|---|---|---|---|---|
| Loading... | |||||
Whitespace Opportunities Donor Intelligence
Largest untapped fundraising potential by candidate and donor
Largest Whitespace by Candidate
Top candidates by total estimated ask from whitespace prospects
| Candidate | Total Ask | Prospects | Avg Score |
|---|---|---|---|
| Loading... | |||
Most In-Demand Major Donors
Donors appearing on 5+ candidate prospect lists
| Donor | Occupation | Candidates | Total Ask |
|---|---|---|---|
| Loading... | |||
Forecast Trust Calibration
Where our forecasts are strongest and where to apply extra scrutiny
Low-Confidence High-Stakes
Races with confident predictions in poorly-calibrated regimes
| Race | Year | Frontrunner | Win % | Trust |
|---|---|---|---|---|
| Loading... | ||||
Highest-Confidence Opportunities
Competitive races in well-calibrated regimes — trust these calls
| Race | Year | Frontrunner | Win % | Trust |
|---|---|---|---|---|
| Loading... | ||||
Data Coverage Gaps
Election cycles with thin or missing data that may affect predictions
| Year | Election Type | Severity | Expected | Actual | Reason |
|---|---|---|---|---|---|
| Loading... | |||||
Office Intelligence Premium
Office-family insights across judicial, house, senate, and statewide races
Loading office intelligence...
Contribution Flow Over Time
Total contribution dollars by year
Race & Candidate Search Predictions
| Year | Office | Candidate | Party | Inc. | Win Probability | Pred. | Actual |
|---|---|---|---|---|---|---|---|
| Enter a candidate name or office to search 17,000+ predictions | |||||||
Donor Lookup Entity Resolution
| Donor | Type | Total $ | Contribs | Filers | City | Employer | Active |
|---|---|---|---|---|---|---|---|
| Search 2.3M+ resolved donors by name | |||||||
Election Stage Coverage Matrix Gaps
Green = solid data • Amber = thin (<50 rows) • Red = missing
| Year | General | R Primary | D Primary | R Runoff | D Runoff |
|---|---|---|---|---|---|
| Loading coverage data... | |||||
Results by Office Family
Fundraising Coverage Audit
| Year | Candidates | Races | Fundraising % |
|---|---|---|---|
| Loading... | |||
Whitespace Call Sheet Donor Intelligence
Ranked prospect lists — who should this candidate call next?
FEC Federal Data 2004–2026
Federal campaign contributions from TX donors across 12 election cycles
Contributions by Cycle
Top Federal Candidates
TX federal candidates ranked by total individual contributions received
| Candidate | Office | District | Party | Contributions | Total Raised |
|---|
Combined Profile Lookup
Search for a candidate to see combined state + federal fundraising
TEC Lobby Overview Lobby Intelligence
Texas Ethics Commission lobby activity, expenditures, and subject matter
Lobby Spending by Year
Top Spenders
| Name | Total Spent | Years |
|---|
Top Subject Areas
| Subject | Reports |
|---|
Dark Money & Independent Expenditures Follow the Money
FEC Schedule E independent expenditures + IRS 990 nonprofits targeting TX races
Support vs. Oppose Spending
IE Spending by Cycle
Top Independent Expenditure Spenders
| Committee | Total Spent | # Expenditures | Candidates |
|---|
Top 990 Nonprofits (501c4/5/6)
| Name | City | Section | Assets | Revenue |
|---|
Local Campaign Finance City-Level
Municipal campaign contributions from Texas cities
Top Local Donors
| Donor | City | Total | Contributions |
|---|
Top Recipients
| Recipient | City | Total | Donors |
|---|
Donor Coalitions Graph Intelligence
12 communities detected via Louvain clustering on shared-donor networks
| Community | Party | Office Focus | Filers | Donors | Total $ |
|---|---|---|---|---|---|
| Loading communities... | |||||
Power Brokers Network Hubs
Donors with outsized reach — bridge donors, kingmakers, and bipartisan operators
| Donor | City | Employer | Filers | Total $ | Community | Flags | Entropy |
|---|---|---|---|---|---|---|---|
| Loading power brokers... | |||||||
Office Coalition Overlap Shared Donors
How much donor base do office families share?
Select an office family above
Trust Center Loading...
How much should you trust our forecasts? Model quality, calibration, and data integrity.
Overall Accuracy
--
Brier Score
--
Lower is better (0 = perfect)
Calibration (ECE)
--
Lower is better
Quality Gates
--
Walk-Forward Backtest Out-of-Sample
Each year predicted using only data from prior elections
Calibration Reliability
Regime Leaderboard
| Regime | Races | Accuracy | Brier | Calibration |
|---|---|---|---|---|
| Loading... | ||||
Confidence vs. Accuracy
| Confidence Band | Predictions | Accuracy |
|---|---|---|
| Loading... | ||
Year-by-Year Detail
| Year | Races | Acc. | Brier | ECE | Trust |
|---|---|---|---|---|---|
| Loading... | |||||
Quality Gates & Provenance Integrity
Promotion Gates
Loading...
Build Provenance
Loading...
How we measure this ▸
Walk-Forward Protocol
Every election year is predicted using a model trained exclusively on data from prior years. The model never sees future outcomes, mimicking real forecasting conditions. This is the gold standard for evaluating political prediction models.
Brier Score
The mean squared difference between predicted probabilities and actual outcomes. Range 0–1, where 0 is perfect. A Brier score below 0.20 indicates strong probabilistic accuracy. Our score of ~0.13 means predictions are well-calibrated.
Expected Calibration Error (ECE)
Measures whether predicted probabilities match observed frequencies. When we say “70% chance,” does that candidate win ~70% of the time? ECE below 0.03 = “Strong,” 0.03–0.08 = “Moderate,” above 0.08 = “Low Confidence.”
Regime Splitting
Texas elections vary widely: a judicial general (party-line voting) differs from a legislative primary (fundraising-driven). We train separate models for ~10 regimes to capture these structural differences. Each regime has its own trust level.
Ensemble Weighting
We blend 4 models (GBM, logistic, regime-split, ensemble) using inverse-ECE weighting: better-calibrated models get more influence. The promoted model must pass 5 quality gates before it replaces the prior version.
Quality Gates
Before promotion, every model must: (1) not degrade calibration, (2) not collapse any regime’s accuracy by >5pp, (3) beat the “pick highest fundraiser” baseline, (4) beat or match the prior model, (5) pass all 10 temporal leakage tests.
Vendor Intelligence SPEND ANALYSIS
Search vendors, analyze market share, and track campaign spending relationships
| VENDOR | CLIENTS | REVENUE | SERVICES | ACTIVE | |
|---|---|---|---|---|---|
| Search 580K+ vendor relationships | |||||
Market Intelligence CONCENTRATION
| VENDOR | CLIENTS | REVENUE | ACTIVE |
|---|---|---|---|
| Select filters and click Load Rankings | |||
Report Builder Brief Studio
Generate client-ready briefs from intelligence data
Race Intelligence
Candidates, predictions, benchmarks, trust context
Candidate Benchmark
Pacing position, projection, fundraising gap
Donor Call Sheet
Ranked whitespace prospects with ask amounts
Donor Profile
Full donor history, recipients, network position
Judicial Race Brief
Plaintiff/defense split, firm concentration, bar alignment
District Digest
Office family overview with competitive races & benchmarks
Batch Generation VOLUME
Generate reports for all races in an office family
Workspace Operating
No items saved yet. Use the bookmark icon on races and donors to add them here.
Recent Activity
No activity yet
Client Management Consulting OS
Loading clients...
Delivery Calendar Schedule
| Client | Product | Cadence | Last Delivery | Next Due | Status | Actions |
|---|---|---|---|---|---|---|
| Click Refresh to load calendar | ||||||
Product Catalog Reference
| Product | Tier | Description | Price Range | Unit | Cadence |
|---|---|---|---|---|---|
| Loading... | |||||
Opportunity Queue Internal
Races ranked by consulting potential — underfunded, competitive, premium
| Score | Race | Year | Type | Raised | Bench P50 | Gap % | Win % | Actions |
|---|---|---|---|---|---|---|---|---|
| Click Search to load opportunities | ||||||||
Pipeline Modules
| Module | Description | Status |
|---|---|---|
| 01 | TEC Data Download | Complete |
| 02 | Silver Layer & Canonical Schema | Complete |
| 03 | Office Ontology & Classification | Complete |
| 04 | Candidate-Filer Linkage v2 | Complete |
| 05 | Donor Entity Resolution | Complete |
| 06 | Election Coverage Completion | Scoped |
| 07 | Dashboard & Visualization | Complete |
| 08 | Candidate Prediction Model | Complete |
| 09-19 | Remaining modules | Planned |
Data Quality Notes
Election data gap: 1996–2000 results missing from SOS scraper
Identity resolution: Name-prefix matching at ~60–70% match rate
Donor deduplication: 5,624 merged clusters with varying confidence
Model regimes: 11 regimes — judicial gen (pre/post-2020), judicial primary/runoff, legislative gen/primary/runoff, statewide, D/R primary, other
Contribution records: 34.8M rows from TEC, spanning 1990–2026
Top Employers by Donor Count
Fundraising Benchmarks Pacing
Is this candidate on pace relative to comparable winners?
Benchmark Reference Bands
Historical fundraising bands by office, election type, and seat status
| Office / Group | Type | Seat | P25 | Median | P75 | Winner Med. | Races | Sample |
|---|---|---|---|---|---|---|---|---|
| Select an office family and click Apply | ||||||||
Candidate vs. Benchmark 0
| Year | Office | Candidate | Party | Raised | Bench P50 | % of P50 | Tier | Gap to Winner | Sample |
|---|---|---|---|---|---|---|---|---|---|
| Click Apply to load candidate benchmark data | |||||||||