Overview 

I led the development of a machine‑sensor analytics dashboard that unified multiple data sources into a single source of truth for machine‑level performance. The first version (2024) gave my department its first real visibility into why defects were happening and how to prevent them, enabling faster and more consistent troubleshooting. In 2025, after the plant upgraded its control systems, I rebuilt the entire solution with richer sensor data, a new database structure, and automated refreshes, transforming it into a scalable analytics tool adopted across multiple sites.

Problem

Before this project, defect analysis was slow reactive, and inconsistent. Data lived across:

• MES logs

• SAP quality results

• Excel files

• Legacy sensor programs

There were no visuals showing which machines, styles, or colors were driving issues. I relied on manual spreadsheets, and operators had no real‑time visibility into machine behavior.

This created:

• Delayed root cause analysis

• Repeated failures

• Inconsistent troubleshooting

• Limited ability to identify trends

The department needed a unified, reliable analytics system.

Aproach

2024 - First Build (Fragmented Data → First Unified View)

Context

In 2024, I independently initiated the first version of the dashboard after realizing that the department lacked a clear way to analyze defects. This was not part of my formal responsibilities, I saw the opportunity and built the solution.

What I Built (2024)

• A Power BI dashboard showing machine‑level KPIs, defect trends, and failure patterns

• SQL logic joining MES logs, SAP quality data, Excel files, and legacy databases

• The first unified view of machine performance for engineering analysis

My Role (2024)

• Built the SQL model and dashboard end‑to‑end

• Created KPIs for defect rate, downtime, and failure patterns

• Trained engineers on how to use the dashboard for investigations

How It Worked (2024)

MES + SAP + Excel → SQL → Power BI (manual refresh)

• Limited sensor visibility

• Inconsistent data structures

• First iteration of drilldowns by machine, shift, and time window

2024 Dashboards:

2025 - Full Rebuild (Adaptive HMI → Clean, Centralized System)

Context

In 2025, the plant migrated to Adaptive, a new HMI/interface with a completely different data structure. This created an opportunity to rebuild the system from the ground up.

I partnered directly with the engineering team responsible for the new platform to design the new database and unlock sensor data we never had before.

What I Built (2025)

• A new SQL model aligned to the updated data structure

• A centralized database consolidating all machine data

• Automated Power BI dashboards with deeper machine insights

• Operator‑friendly pages for real‑time decision‑making

My Role (2025)

• Co‑designed the new database structure with platform engineers 

• Rebuilt all SQL logic, DAX, and Power Query transformations

• Created second‑generation dashboards

• Trained both engineers and operators on the new system

How It Worked (2025)

New Database → SQL → Power BI (automated refresh)

• Clean timestamps

• Richer sensor tags

• Faster refresh cycles

• More accurate KPIs and trend analysis

2025 Dashboards:

Results

• ~50% reduction in machine‑driven defects 

• More than $300K in annualized savings enabled by faster root‑cause analysis

• Recognized with a 4th Quarter Milliken Performance System award for enabling data‑driven defect reduction

• Replicated at additional sites due to its scalable design

• Improved engineering analysis and operator decision‑making

📌 Add Visual Here:

Before vs. After defect chart (blur internal identifiers)

Key Skills Demonstrated 

• SQL data modeling

• Power BI development

• Multi‑system data integration

• KPI design & visualization

• Root cause analysis support

• Automated reporting pipelines

• Cross‑functional collaboration

• Operator‑focused design