Top Brief Calibration & Export

Weights auto-normalize to 100%. Win threshold is percentage points. Volatility is coefficient of variation for booked revenue.

90-Day Booked Revenue Forecast by Estimator

Forecast uses estimator-level monthly history with lag features and seasonality. Error bars show an approximate 80% interval from in-sample model error.

Executive Action Summary

Composite score combines win quality, efficiency, stability, and expected near-term revenue.

Estimator Performance Scorecard

Data Science Insight: Pipeline Correlation

Annual Performance

K-Means Behavioral Segmentation (3-Feature Model)

Clusters on: Pipeline Effort (Total Bid $), Efficiency (Lag-Adjusted Win Rate %), Selectivity (Avg Job Size $). Point size = Total Booked. Hover for details.

Estimator Leaderboard — Click any column to sort

Intervention Settings


The date a strategy change, new hire, process shift, or other intervention occurred.


Uses the strucchange algorithm to scan for the single most statistically significant structural break in the series.


Intervention Effect — Statistical Summary

Segmented regression (OLS) with pre/post intervention terms. Level change = immediate shift at intervention date. Slope change = difference in monthly trajectory before vs. after.

Interrupted Time Series — Visual

Grey points = actual monthly values. Blue line = pre-intervention fitted trend. Green line = post-intervention fitted trend. Vertical dashed line = intervention point. Red dotted = counterfactual.

Auto-Detected Structural Breaks

All significant breakpoints detected by the strucchange algorithm, ranked by F-statistic. Click a row to set it as the active intervention date.

Coefficient Detail

Full regression output. p < 0.05 indicates a statistically significant effect.

Upload Data

System Status


                    

Export Report

Download a formatted Excel workbook of the current filtered view.

Download Excel Report

Date Verification — Raw vs. Parsed

Confirms MoYear values are parsed correctly. Format: YY-MM (e.g. '19-06' → June 2019).

Experimental Tab — Powered by IMS bid follow-up records. Shows contractor relationships, pipeline activity, and award signals extracted from estimator field notes. Data source: bid___follow_up joined to estimators .

Filters

to

Award Signal Distribution by Estimator — NLP Scored

Each note scored by a Logistic Regression model trained on 2,454 follow-up records . Colors indicate award signal confidence. Run nlp_followup_scorer.py to refresh scores.

Top Contractors by Follow-Up Activity

Bars show total follow-up contacts per GC across all selected estimators. 'Estimators' mode shows how many different estimators have contacted each GC.

Follow-Up Activity Timeline

Monthly or quarterly count of follow-up contacts logged. Reveals pipeline activity intensity over time.

Estimator × Contractor Relationship Matrix

Heatmap of follow-up contacts. Darker = stronger relationship. Reveals which estimators own which GC relationships.

Award Signal Feed — Notes containing win/award language

Filtered for notes containing: awarded, booked, proceed, LOI, notice to proceed, doing this project, re-awarded

Award signals extracted from IMS follow-up notes. Use as a win log proxy until full bid outcome data is available.

Phase 2 — Revenue Intelligence. Powered by executed project data from the IMS PM system. Shows contract value, project completion, and GC revenue concentration by project manager. Source: pm___project_file joined to project_file .

Filters

to

Executed Contract Value by Project Manager

Total contract value of executed projects per manager. Source: pm___project_file joined to project_file on projectid.

Project Completion Pipeline

Completed vs in-progress projects per manager. Toggle to see by count or by contract value.

GC Revenue Concentration

Which GCs generate the most executed contract value. Reveals revenue concentration risk per contractor relationship.

Executed Project Detail

Full project list with contract values, completion dates, and GC. Sortable and searchable.