Dr. Data Mining — Industrial AI That Gets Your Team to Act

News

SME Annual Meeting — Salt Lake City

Dr Data Mining to Present Core Natural Resources Digital Transformation Results

Dr. Sean Dessureault will co-present with Core Natural Resources' Eric Shereda during the COAL & ENERGY: TECHNOLOGY, INNOVATIONS, & AI/ML session on Tuesday, February 25th. The presentation highlights how operational data, time-series intelligence, and performance engagement workflows can be deployed in real mining environments to support measurable operational impact.

Enterprise Expansion

Phase 2 Deployment Underway With Core Natural Resources

Following promising Phase 1 results at the Bailey Coal Preparation Plant and across the Pennsylvania Mining Complex, Core Natural Resources has elected to expand the program into Phase 2 — scaling TSAI and Performa from Pennsylvania to Core's broader portfolio of underground mines and preparation plants.

The Real Problem

Most AI projects
never make it
past the dashboard.

The algorithms work. The data is there. The dashboard looks great on the screen in the boardroom. But six months later nothing has changed on the plant floor. We've watched this happen for thirty years. The gap isn't in the math — it's in the operationalization.

Insight Without Accountability

Anomaly detected at 2am. Nobody on shift knows what it means or what to do. The alert fires, gets ignored, and the pattern repeats.

Models That Drift and Die

A consultant builds a great model. It works for 90 days. Then the process changes, the model goes stale, and nobody owns the retraining. Cost: the entire engagement.

Operators Who Don't Buy In

Blue-collar workers don't change behavior because of a PowerPoint. They change behavior when they can see their own performance, understand why it matters, and have skin in the game.

Data That Isn't Actually Connected

Your SCADA, your ERP, your shift logs, and your lab results have never been joined in one place. Every optimization you've tried was working with incomplete information.

Our Approach

We Close the Gap Between
Intelligence and Action

Most AI projects stall between the model and the plant floor. We've spent 30 years in that gap. Our engagements start with a consulting phase — understanding your operation, your data, and your people before recommending anything. If a platform isn't the right answer for you, we'll tell you that.

Consulting & Assessment

We assess your data landscape, technology stack, and operational maturity. You get an honest picture of what's possible — not a sales deck. We've helped clients decide against projects that wouldn't have paid off. That's how we build relationships that last decades.

Data Exploration & AI Development

We connect to your data, explore it systematically, and develop AI algorithms tailored to your process. Statistical process control, operator profiling, blend optimization, anomaly detection, root cause analysis — each finding is backed by your own numbers, not generic benchmarks.

Operationalization & Platforms

Findings don't pay the bills — changed behavior does. We build the real-time data infrastructure, deploy the dashboards operators actually use, and tie performance to accountability through gamification. The software is the delivery mechanism for the insight.

See Our Full Methodology →

≠

Honest AI Consulting — Including the Hard Truths

We've seen the AI hype cycle up close — we were doing this before it was cool. Here's what we tell clients that most vendors won't: AI projects fail for exactly two reasons — inadequate data infrastructure, and organizations that don't follow through on what the model finds. Most vendors solve neither. We've built platforms specifically for both: TSAI organizes and contextualizes your operational data so the algorithms actually work, and Performa makes sure your team acts on what they find. That's why our deployments stick.

Real Work. Real Results.

In partnership with Core Natural Resources — deploying TSAI and Performa across one of the largest coal producers in the United States, scaling AI-driven process intelligence from initial deployment to enterprise-wide operations.

Coal Preparation Plant
Analytics — Active Deployment

One of the world's largest coal preparation plants. Three mines. Five longwalls. Thousands of tags. Here's what the data actually showed.

Core Natural Resources
US Coal Producer
Active 2025–2026

26.7%

Unnecessary Magnetite Additions

Over a quarter of magnetite additions to the heavy-media circuits were flagged as unnecessary or counterproductive — operator-driven, not process-driven. Magnetite is expensive. This is a direct, quantified cost reduction opportunity.

2×

Chute Plug Rate — Plant 2 vs Plant 1

1,492 chute plug events analyzed. Plant 2 shows 2× the plug rate of Plant 1. Root cause: Enlow Fork coal's clay content, not operator error or abnormal blend ratios. A finding that changes the operating procedure, not just the alert threshold.

Distinct Operator Behavioral Clusters

K-means clustering on operator response patterns identified 4 distinct profiles. One profile consistently shows least magnetite usage but worst SG control — a counter-intuitive finding that conventional metrics would miss entirely.

29%

Shift Changes With Mag Dump Behavior

Of 42 shift transitions analyzed, 29% showed a "mag dump" — unnecessary magnetite addition immediately after handover. A behavioral pattern invisible to conventional reporting, now quantified and addressable.

Analysis Methodology — This Engagement

Data Connection (QuestDB) → SPC & Drift Classification → Operator Profiling → Blend Optimization Models → Root Cause Attribution → Operator Scorecards → Deployed Dashboards

Operations Timeline — Plant Status

Live TSAI Dashboard — Silo Inventory

TSAI Live Inventory Dashboard — Bailey Prep Plant silo levels

Operator Profiling — Performance Scorecard

Operator Performance Scorecard — per-operator behavioral analysis

MSHA Benchmark AI — Peer Comparison

MSHA Peer Comparison — TRIR and DART rate trends across companies

The Platforms

Built for Mining.
Deployable in Days.

Two platforms that handle the full operationalization cycle — real-time intelligence from the plant floor to actionable performance for every operator on every shift.

Time Series AI Platform

TSAI

From sensor to insight in real time.

TSAI is the data infrastructure and analytics engine behind every deployment. It connects to your existing automation systems — OPC-UA, MQTT, PLCs, SCADA — and builds a unified time-series foundation for AI analysis. No rip-and-replace. No IT project. Running in Docker on an on-premise server at your site.

Coal Magnetite Effectiveness — Real-time SG monitoring, unnecessary addition detection, shift reporting

Blend Optimization — Linear regression, Random Forest + SHAP, constrained optimization for max throughput within quality specs

Process Health Monitoring — Diverter valve diagnostics, mechanical anomaly detection, degradation trending

OEE Analytics — Equipment effectiveness, delay classification, availability and utilization tracking

Plant Timeline — Gantt-based visualization integrating SCADA, delays, blend, and quality on one screen

LLM Shift Reports — Natural language summaries of plant performance generated from analysis data

Explore TSAI →

Performance Gamification Platform

PERFORMA

The missing link between insight and behavior change.

The reason AI projects fail isn't the algorithm — it's that nobody changed what the operators do on the plant floor. Performa closes that gap. It takes the KPIs and recommendations from TSAI and turns them into a real-time performance system with scores, rankings, and shift-level accountability that resonates with blue-collar workers.

Operator Scorecards — Per-operator, per-shift performance scores based on actual process behavior

Team Rankings & Leagues — Shift vs shift, crew vs crew — competitive structures that drive engagement

Goal Tracking — Daily, weekly, and monthly targets tied to throughput, yield, and quality KPIs

Achievement System — Milestones and recognition for sustained performance improvement

Management Dashboard — Site-wide view across all shifts and circuits with drill-down to individual operators

Excel Add-In — Full bulk data management for supervisors and data administrators

Explore Performa →

Safety Intelligence

MSHA
Benchmark AI

Every US mining company files detailed safety data with MSHA — violations, accidents, inspections, production hours, employment, contractor activity, and more. Decades of mandatory reporting across the entire industry have made this one of the most researched and studied mining safety datasets in the world, with a deep body of peer-reviewed academic work built on top of it. The question is: how does your company take advantage of that mountain of research and data to benchmark your own performance?

MSHA Benchmark AI gives you three capabilities. Internal benchmarking — compare your own sites against each other to find outliers and best practices. Peer benchmarking — measure your performance against every operator in your commodity and region. And research-grade analysis — we use LLMs to systematically read and apply peer-reviewed academic research on mining safety and injury rates, under the supervision and curation of a human expert, to build deep analyses against your own data combined with the full MSHA public dataset. Your operation measured against published science, not just against your competitors.

The findings don't stop at a report. Each analysis produces action-oriented recommendations that feed directly into Performa's accountability system — turning research insights into specific, trackable follow-through on the ground.

Updated weekly from the MSHA open data portal. AI anomaly detection flags unusual spikes in violations or injuries. Predictive risk scoring identifies mines trending toward Pattern of Violations status before MSHA does. Natural language executive summaries included.

Learn More →Request a Report

20+

MSHA Datasets — Violations, Accidents, Inspections, Production

Since 2000

Historical data from the MSHA open portal, updated weekly

Internal
+
Peer

Benchmark your own sites against each other and against industry peers

POV

Early warning risk scoring — identify Pattern of Violations trends before regulators do

Peer Comparison Pool

Peabody Energy · Arch Resources · Alpha Natural Resources · Alliance Resource Partners · Warrior Met Coal · Foresight Energy · CONSOL Energy — and any operator in the MSHA registry.

Accessibility

Your Team Already Knows Excel

Both platforms are fully accessible from Microsoft Excel via C# add-ins built with Excel-DNA — no new software to learn, no browser required, no IT approval needed for most operations. Data administrators and supervisors work in the tool they already use every day.

TSAI Excel Add-In

Query time-series data directly from QuestDB into Excel. Pull any tag, any time range, any aggregation — without writing SQL. Bulk tag management, UNS namespace administration, and site configuration across all 11 deployments.

Performa Excel Add-In

Bulk data entry and management for operator records, KPI targets, scoring rules, and gamification configuration. Three-action patterns (Load / Update / Delete) for all lookup tables. Role-based access control built in.

Visualization & Display

Dashboards That Belong
on the Plant Floor

Industrial displays aren't web apps. They run 24/7 on screens in control rooms, on the floor, and in management offices. We build for that environment — screen rotation systems that cycle through HMI emulations, gamification boards, KPI dashboards, and embedded Power BI reports in a single continuous loop.

HMI Emulation

We recreate HMI-style process screens in React — the same visual language operators already know from their SCADA systems, but with analytics overlays, anomaly highlights, and real-time AI recommendations embedded directly into the familiar interface. No new mental model required.

Screen Rotation

Control room displays rotate automatically between content types — process health dashboards, operator performance scoreboards from Performa, HMI screens, embedded Power BI reports, and shift summary cards. One screen. Full operational picture. No manual switching.

Power BI Integration

If your organization already runs Power BI reports, we embed them directly into the rotation alongside TSAI and Performa displays. Existing BI investment preserved. Data warehouse reports and real-time sensor dashboards coexist on the same screen.

What Rotates on a Typical Display

Plant process overview — HMI-style with live sensor values

Performa scoreboard — Current shift operator rankings

KPI summary card — Throughput, yield, quality vs targets

Active alerts — SPC violations, anomalies, maintenance flags

Power BI report — Embedded from your existing BI environment

Shift handover summary — AI-generated plain-language recap

Built For the Industrial Environment

Designed for large-format displays in dusty, high-ambient-light environments. High contrast, large typography, color-coded status. Runs on standard industrial PCs and thin clients. No cloud dependency — everything operates on-premise on your site network.

About

30 Years of Mining AI.
Before It Was Cool.

Dr. Data Mining was built around a simple observation: the mining industry has never lacked for data. It has lacked for people who could turn that data into decisions that stick on the plant floor. Founded by Dr. Sean Dessureault — a mining engineer, serial entrepreneur, and former tenured professor — the company brings together three decades of experience designing, deploying, and supporting data systems in heavy industry.

That experience spans the full cycle. From building the first university research program in mining information technology at the University of Arizona, to founding MISOM Technologies — a mobile app and IoT company that was bootstrapped from consulting, grew through a Series A funding round, and was acquired by MST Global (now Komatsu) in 2017. From serving as VP of Technology and Innovation at The Mosaic Company, scouting and evaluating emerging technology across global operations in Florida, Saskatchewan, Brazil, Saudi Arabia, and Peru — to building Dr. Data Mining as the culmination of everything learned along the way.

Our clients typically engage us first as consultants. We assess their data maturity, identify the highest-value optimization opportunities, and develop AI algorithms against their actual process data. When the findings are compelling enough to warrant a permanent system — which they usually are — we deploy TSAI and Performa as the operationalization layer. The consulting engagement funds itself through the performance improvement it uncovers.

Currently deploying across 11 active sites at a major US coal producer — prep plants and mines across four states.

We live comfortably in both the time-series world and the traditional data warehouse world. We design and deploy SQL Server and PostgreSQL data warehouses, SSAS tabular models, and integrated datasets that bring together SCADA, ERP, lab, and shift data. TSAI handles the real-time sensor layer. The data warehouse handles everything else. They work together.

Our Philosophy

Start Small

Few large initiatives succeed. Small seeds grow strong roots.

Use What Works

We have an excellent legacy of analytics. Use known standards.

Follow Through ASAP

Flow through the entire data-to-action process as fast as possible. Start using data.

Use Brings Accuracy

The most used data is the most accurate. Don't wait for perfection before acting.

Engage the Workforce

Through detailed feedback, they'll value the data more.

Avoid the Hype

Consultants and tech companies often overhype. Rely on what works.

Founder

Dr. Sean Dessureault, Ph.D. — Mining Engineer, Serial Entrepreneur, Former Tenured Professor

Experience

30+ years in industrial data analytics, mining AI, and operations intelligence

Recognition

American Mining Hall of Fame — "Medal of Merit under 40" · TEDx Speaker · Chair of International APCOM Council

Prior Roles

VP Technology & Innovation, Mosaic Company · Chief Innovation Evangelist, MST Global · Founder & CEO, MISOM Technologies (acquired by Komatsu) · Associate Professor, University of Arizona

Education

Ph.D. & M.Sc., Mining Engineering — University of British Columbia · B.Eng., Mining Engineering — McGill University

Active Deployments

11 sites — PA, WV, CO, WY

Commodities

Coal preparation · Metal mining · Mineral processing · Phosphate & potash · Industrial operations

Time-Series Stack

QuestDB · Node-RED · OPC-UA · MQTT · Redpanda · FastAPI · React · Docker

Warehouse Stack

SQL Server · SSAS Tabular · PostgreSQL · PowerBI · Excel-DNA · C#

Research

$4.5M+ in funded research · PhD and Master's student supervision · Published extensively in mining technology and sustainability

Current Focus

LLM integration for shift reporting, MSHA Benchmark AI, Process Health Monitoring

ANY PLATFORM CANSHOW YOU THE DATA.WE GET YOUR TEAMTO ACT ON IT.

Analysis Methodology — This Engagement

ANY PLATFORM CAN
SHOW YOU THE DATA.
WE GET YOUR TEAM
TO ACT ON IT.