EnviroTech Dynamics – NumPy Temperature Analysis

A fast, vectorized data-analysis project built using NumPy to process 1 million+ temperature readings and detect anomalies without loops.

🚀 Project Summary

This project replaces slow, loop-based inspection with fast, vectorized NumPy operations. It analyzes environmental temperature data, identifies critical anomalies, and cleans faulty sensor readings using statistical techniques.

📁 Dataset Structure

• Temperature readings: 1,000,000 values
• Status codes (system health): 0 → Normal

1 → Warning

2 → Critical

3 → Sensor Error

🧠 Key Features

• Fast statistical analysis (mean, median, std, percentiles).
• Boolean masking for instant anomaly detection.
• Vectorization & broadcasting for complex operations.
• Faulty data replacement using median imputation.
• Matplotlib visualizations for insights.
• Performance checked with %timeit.

⚠️ Critical Findings

• 34 critical temperature spikes found.
• Values ranged 81°C–86°C, far above normal levels.
• Only 0.0034% of all readings, but high-risk.

🧼 Data Cleaning Results

• 20,102 faulty records (status = 3) replaced with median (~45°C).
• Median remained unchanged (44.99°C) after cleaning.
• Dataset quality remained stable and reliable.

📊 Visual Outputs

• Plot of temperature distribution.
• Highlighted anomaly visualization.

🏁 Final Insights

• NumPy reduced analysis time from minutes to milliseconds.
• Loop-free vectorization made the system scalable.
• Data cleaning improved reliability without changing distribution.
• The pipeline is now optimized, efficient, and ready for real deployment