Spend Analytics

Dashboard

🧭 Guided Workflow

Spend by Region

Spend by Category

Expiring Contracts

Top 10 Vendors by Spend

Contract Status Distribution

Spend Under vs Without Contract

All Contracts

Full contract inventory with contract end-date visibility and quick filters

IDContract NameSupplierStatusCategoryClusterRegionEnd DateDays to EndSpend (USD)Confidence

System Wiring

Deep map of how files, APIs, pipeline stages, and screens connect in this build.

Status not loaded yet

Pipeline Stages

Screen & API Coupling

Artifacts & Persistence

Runtime Health

Observability (Palantir-style)

Pipeline Runs

Who ran what and when; stage durations and status.

Run IDStartedDurationStatusStagesArtifacts

Data Lineage

Last run and asset → run mapping.

Asset Graph

Nodes: runs and artifacts; edges: produced/consumed.

Contract Hierarchy

Parent-child relationships, amendments, and vendor groupings

Parent IDParent ContractChild ContractSupplierRelationshipConfidence

Spend Analysis

Region x Category spend breakdown

Spend by Region x Category

Spend by Spend Type

Contract Value vs Actual Spend

Contract Clusters

Unsupervised segments by spend, region, commodity, and contract value (K-means)

Spend by Cluster

Cluster Summary

ClusterContractsTotal Spend (USD)

Contract Connection Graph

Interactive XYFlow map of contract links, grouped by predicted category (functional/logistics/etc.).

Ready
MSA node SOW node Other node Approved edge Needs Review edge Rejected edge

XYFlow Connection Map

Loading connection graph...

Expiring & Negotiated Contracts

Visual focus on contract end dates by category with search and filters

Expirations by Year

Negotiations by Year

Expiring by Category (Next 12 Months)

Contract End Date Queue

IDContractSupplierEnd DateDays to EndStatusSpendRegionCategory

Vendor Concentration

Spend concentration and vendor analysis

Spend Concentration (Pareto)

Vendor Count by Category

Vendor Summary

SupplierContractsTotal Spend% of TotalCumulative %RegionsCategories

Reinforcement Learning Loop

Model training status, Q-learning metrics, and active learning progress

📈 RL Agent Status

Model Confidence Over Rounds

Epsilon Decay (Exploration vs Exploitation)

Cumulative Reward

Corrections per Round

How the RL Loop Works

1
Predict
ML model classifies each contract into a category
2
Sample
Active learning picks lowest-confidence items for review
3
Label
Category managers confirm or correct predictions
4
Reward
+1 for correct, -1 for wrong. Q-values updated
5
Retrain
Model retrains, epsilon decays, policy improves

Input File & Build Model

Add a file for processing, then run the pipeline to build or rebuild the model (cleaning, linking, clustering, fitting).

📄 Current input file

Loading...

Upload new file (Excel)

Upload an Excel file with contract/spend data. It will be set as the current input. Then click "Run pipeline" to process and build the model.

Run pipeline / Build model

Run the full pipeline: clean data, link contracts (including MSA→SOW), classify vendors (ML), analytics, then refresh report and dashboard. Use after uploading a new file or after attaching MSA→SOW.

Monthly Operating Cycle

Persistent monthly stage/status tracking and backward remediation proposals when source systems cannot be changed immediately.

Fix Proposals (Top 200)

IDPriorityIssueEntityStageStatusProposed FixOwner

MSA → SOW Attachments

Attach multiple Statement of Work (SOW) contracts to one Master Service Agreement (MSA). Run pipeline after saving to refresh hierarchy.

Attach SOWs to an MSA

Select one MSA (contract) and one or more SOW contracts to link. You can attach many SOWs to a single MSA.

Current MSA → SOW links

MSA (Contract ID)MSA RoleContract nameAttached SOWs (ID / role)

Label & Teach the Model

Review AI predictions and correct them. Sorted by uncertainty (most uncertain first).

Flags & Alerts

Automated risk detection and contract issues

SeverityFlagContract IDSupplierDetail