"Dad built it with his hands. We'll keep it growing with data."
This is a structured, evidence-based proposal for how Kedren Reade Sitton — 25+ years in SAS ETL, data engineering, and analytics platform architecture — can step into Northstar Electric and create measurable, lasting value for a family business entering its next chapter.
It is not a résumé drop. It is a north star: a fixed navigational point that answers the question "How exactly does a data engineer help an electrical contracting and retail electric services company?" — with specifics, not generalities.
Northstar Electric is a full-service electrical contracting and retail electric services company with decades of operational history. It does two things that generate two very different kinds of data:
| Division | Core Activity | Data Generated |
|---|---|---|
| Electrical Contracting | Commercial/residential install, service, repair | Job costs, labor hours, materials, billing, scheduling |
| Retail Electric Services | Electricity supply/brokerage to customers | Usage data, rate data, customer accounts, margin |
Both divisions are sitting on data they almost certainly aren't fully using. That's the opportunity.
A business that cannot see itself clearly cannot steer.
Northstar Electric has likely run on experience, relationships, and gut instinct — all of which got it here after 40+ years. But the next decade of electrical contracting looks different: energy deregulation complexity, labor cost pressure, materials volatility, and customers who comparison-shop retail electricity rates online.
Data engineering doesn't replace the founder's instincts. It amplifies them — by turning scattered invoices, job tickets, usage files, and QuickBooks exports into a clear, always-current picture of what the business is actually doing.
Here's how 25+ years of enterprise-grade data work translates directly to Northstar Electric problems:
The Problem: Electrical contractors routinely carry 60–120 days of unpaid invoices without a clear picture of total exposure. Cash is sitting in someone else's bank account. The owner knows certain customers are slow payers — but probably doesn't know the aggregate number.
The Reade Solution:
- ETL pipeline: QuickBooks invoice export → aging buckets (Current / 31–60 / 61–90 / 90+)
- Chronic slow-pay customer identification: flag anyone with 2+ invoices over 60 days
- Weekly A/R summary delivering the one number that matters: total outstanding over 90 days
Skills Applied: SQL/ETL, Python/Pandas, automated reporting, cash-flow discipline applied across financial services clients
The Problem: Electricians must maintain state licenses, OSHA certifications, arc flash training, and CDL renewals. A lapsed cert means failed inspections, voided insurance claims, or state fines. Most small contractors track this in someone's head or on a sticky note.
The Reade Solution:
- Database of every employee credential with expiration date
- Tiered alert system: EXPIRED / RED (30 days) / AMBER (60 days) / YELLOW (90 days)
- Automated weekly email: "Two certifications expired. Eight expire this month. Here's the list."
Skills Applied: Python, scheduled automation, the same SLA-compliance discipline from 99%+ uptime work at Optum and Delta Dental
The Problem: Electrical contractors often win jobs based on estimated labor and materials, but rarely close the loop on whether the job was actually profitable after actuals.
The Reade Solution:
- ETL pipeline: job estimate data → actual hours/materials → margin report
- Automated weekly job profitability summary (the same C-suite summary format invented at CrossUSA and used ever since)
- Flag jobs where actuals exceed estimate by >10% for RCA
Skills Applied: SAS/SQL ETL, Python/Pandas pipelines, automated reporting, RCA discipline (Coastal Community Bank), change management
The Problem: Retail electric services means managing customer accounts with usage profiles, contract expirations, rate changes, and churn risk — typically tracked in spreadsheets.
The Reade Solution:
- Customer data model: account → usage history → contract terms → renewal dates
- Churn risk scoring: identify customers approaching contract end with rising usage (i.e., most at risk and most valuable to retain)
- Automated renewal alert pipeline — no customer slips through uncontacted
Skills Applied: MySQL/MSSQL, Python pipelines, time-series analysis (CrossUSA marketing mix modeling), machine learning (LSTM, scikit-learn from MIT cert)
The Problem: Field crews are scheduled manually or in basic tools. Visibility into utilization, overtime, and idle time is limited or nonexistent.
The Reade Solution:
- Ingest scheduling/dispatch data → calculate utilization rate per technician
- Flag overtime creep before it hits payroll
- Power BI dashboard: weekly crew utilization, job completion rate, rework jobs
Skills Applied: Power BI (Delta Dental, Coastal Community Bank), SSRS/SSMS, automated reporting at Optum (800-user scale, 99% availability)
The Problem: Electrical materials pricing is volatile. Without tracking, the company can't know which suppliers are most reliable, whether it's getting competitive pricing, or when to buy ahead.
The Reade Solution:
- Purchase order ETL: vendor → material → price per unit → over time
- Price trend dashboards with alert thresholds
- Vendor scorecard: on-time delivery, price variance, quality flags
Skills Applied: ETL pipeline design, time-series trend analysis, SQL aggregations
The Problem: Retail electric resellers need to track wholesale/retail rate spreads to protect margins and advise customers.
The Reade Solution:
- Real-time rate ingestion pipeline using REST APIs + Apache Kafka or NiFi
- Margin dashboard: current buy rate vs. customer sell rate by account tier
- Alert when spread compresses below threshold
Skills Applied: Apache Kafka, NiFi, REST APIs, IoT/MQTT (MIT cert), real-time streaming pipelines
The Problem: The owner has all the knowledge in his head. When he steps back, that knowledge doesn't transfer automatically.
The Reade Solution:
- Single-page executive dashboard (the same format invented in 2011 at CrossUSA and still in use today)
- Pulls from every data source: jobs, crews, customers, rates, vendors
- Delivered automatically every Monday morning, no login required
- Designed to be handed off — documented, version-controlled, reproducible
Skills Applied: C-suite summary report design (invented and used across CrossUSA, Wells Fargo, Optum, Coastal Community Bank), Power BI, automated delivery, SOP documentation
northstar-electric/
│
├── README.md ← You are here ✦
│
├── docs/
│ ├── value-proposition.md ← Deep-dive narrative proposal
│ ├── skill-crosswalk.md ← Résumé skills mapped to NE problems
│ └── roadmap.md ← Phased 90-day engagement plan
│
├── scripts/ ← All runnable pipelines
│ ├── ar_aging.py ← 🏆 A/R aging + chronic slow-pay flags
│ ├── cert_tracker.py ← 🏆 Certification/license expiration alerts
│ ├── job_cost_etl.py ← 🏆 Job estimate vs. actuals + margin
│ ├── customer_churn_score.py ← Retail electric renewal risk scoring
│ └── rate_alert_pipeline.py ← Rate spread monitor (Phase 3)
│
├── data/ ← Sample data (replace with real exports)
│ ├── ar/ ← accounts_receivable.csv
│ ├── estimates/ ← job_estimates.csv
│ ├── actuals/ ← job_actuals.csv
│ ├── certifications/ ← employee_certifications.csv
│ ├── customers/ ← retail_customers.csv
│ ├── usage/ ← usage_history.csv (12 months)
│ ├── rates/ ← wholesale_rates.csv (30 days, hourly)
│ └── scheduling/ ← crew_schedule.csv
│
├── dashboards/
│ ├── owners-dashboard-mockup.md ← Wireframe of the Monday morning report
│ └── power-bi-notes.md ← Notes on standing up Power BI for SMB
│
└── .github/
└── CONTRIBUTING.md ← For when the team grows
| Phase | Days | Deliverable |
|---|---|---|
| Listen | 1–30 | Shadow operations. Map every data source. Interview owner. Produce the "Data Inventory" document. |
| Light | 31–60 | Stand up the Owner's Dashboard v1. Automate one report the owner currently builds manually. |
| Navigate | 61–90 | Job costing pipeline live. Customer renewal alerts live. First RCA on a job that went over budget. |
Kedren Reade Sitton Senior Data Engineer | MIT xPRO Certified | 25+ Years ETL/Analytics 📧 [email protected] 🔗 linkedin.com/in/reades 🐙 github.com/readesie
You found it. Good. That means you read carefully — which is exactly the kind of person Northstar Electric needs.
Type this in your terminal and see what happens:
echo "Q2hlY2sgeW91ciBqb2IgY29zdHMu" | base64 --decode
The north star doesn't move. But the ships that find it do.
This repository was built as a living document. It will grow as the engagement grows. Last updated: April 2026.