Rian Insights

Data analytics for connected devices — fast, practical, and production-ready

Telemetry → pipelines → dashboards → insights
Helping hardware and IoT teams turn device data into decisions

Services

Analytics & Dashboards — KPI design, realtime dashboards (Power BI / Tableau / Quicksight)
Data Pipelines — SQL/Python ETL, cloud storage & ingestion (S3, Redshift, etc.)
Performance Insights — anomaly detection, reliability scoring, predictive maintenance
Retainers & Fractional Data Leadership — ongoing monitoring, reporting, and roadmap

Fleet Performance & Telemetry Diagnostics for Distributed Energy Devices

Transforming messy device data into actionable operational insights.

Rian Insights analyzed 30 days of telemetry from 60 distributed smart energy devices across the Midwest.
This project highlights an end-to-end workflow: data cleaning, validation, anomaly detection, KPI analysis, and dashboard development — all using realistic, messy device data.

The Challenge

The incoming telemetry contained a wide range of issues commonly seen in real hardware and IoT environments:

  • Missing values (~8%)

  • Duplicate records

  • Mixed timestamp formats

  • Device IDs in inconsistent casing

  • Voltage values stored as strings

  • High-temperature and high-power outliers simulating sensor drift

  • Random jitter and ingestion disorder

Leadership needed a trustworthy, consolidated view of fleet performance to monitor device health, identify risks, and understand operational patterns.


Our Approach

1. Data Cleaning & Validation

  • Standardized device IDs and timestamp formats

  • Removed duplicates and corrected voltage values

  • Interpolated small gaps; flagged large missing segments

  • Identified extreme outliers without removing them

2. Exploratory Analysis

  • Calculated fleet KPIs: power, voltage, temperature, efficiency

  • Analyzed diurnal patterns and regional differences

  • Visualized variability across device models

3. Dashboard Development

  • High-risk device identification

  • Temperature cycles

  • Daily energy trends

  • Regional performance comparisons

  • Status distribution (OK/WARN/ERROR)


Key Insights

  • Efficiency drops at higher temperatures (above ~30°C).

  • Outlier clusters revealed faulty sensors needing attention.

  • Regional patterns in device performance were identified.

  • Devices with frequent missing data or duplicate rows were flagged.

  • Voltage type errors were corrected without losing any historical readings.

Deliverables & Impact

  • Cleaned Telemetry Dataset: Fully validated, missing values addressed, outliers flagged.

  • Merged Analytics Dataset: Combined telemetry with device metadata for richer insights.

  • Interactive Dashboard: Fleet KPIs, trendlines, regional comparisons, and high-risk device tracking.

  • Technical Report: Documenting data issues, cleaning steps, analysis methods, and actionable recommendations.

Results:

  • Data reliability improved from ~92% → ~98%

  • Duplicate records (~2%) removed

  • Hundreds of anomalies flagged for review

  • Executive-ready dashboard delivered for actionable insights

Tools used: Python, Pandas, NumPy, SQL, Jupyter, Power BI / Tableau / Plotly, Scikit-learn

Interested in turning your device telemetry into actionable insights?
Contact Rian Insights today to discuss how we can help your hardware or IoT projects generate real operational value.

Projects & Insights

Rian Insights transforms raw device and operational data into clear, actionable intelligence.
This portfolio highlights real-world analytics projects demonstrating data cleaning, anomaly detection, and fleet performance insights across distributed hardware systems.

Request a free consult