Sensor Location: Maypangdan, Borongan City Online View Map ↗

Air Quality Index 5-MIN INTERVAL

000 AQI INDEX Loading Data

0 50 100 150 200 300+

Outdoors Permitted

Ventilation Open Windows

Sensitive Low Risk

Facemask Not Required

Analyzing atmospheric conditions...

24-Hour Average: —

Standard: DENR DAO 2000-81

3-Hour Predictive Outlook Random Forest ML

+1 Hour

—

━ 0

+2 Hours

—

━ 0

+3 Hours

—

━ 0

Predictive Trajectory Curve: Steady Path

Random Forest Trend

Stable air quality predicted over the next 3 hours.

Environmental Telemetry

Real-time localized ambient microclimate metrics

Temperature

00.0 °C

Ambient Temperature

━ Steady

Humidity DHT22

00.0 % RH

Relative Humidity

━ Steady

Particulate PM10

00.0 µg/m³

Laser Scattering (PMS5003)

━ Steady

Gas Contaminants

000 ADC Index

Relative Baseline

━ Steady

Looking for In-Depth Diagnostics? Inspect model validation benchmarks (R² = 0.952), Isolation Forest anomaly Z-scores, and interactive sensor history timelines.

Daily Summary & Comparisons loading…

Daily Summary Methodology

Method: Database aggregation & comparisons.
Function: Compares today's average, minimum, and maximum readings against yesterday's.

Update Cadence: Aggregated and refreshed automatically every 60 seconds.

Today's Avg

—

AQI

Yesterday's Avg

—

AQI

7-Day Weekly Avg

—

AQI

30-Day Monthly Avg

—

AQI

Min AQI —

Max AQI —

Readings —

—

Interactive Sensor History Timeline —

Key Prediction Drivers (Random Forest)

Random Forest Feature Importance

Gini-importance (mean decrease in impurity) computed across all 100 decision trees to determine which environmental parameters exert the strongest influence on predicted AQI.

Model Evaluation Benchmark R² = 0.952

Architecture	MAE	R² Score	Status
Random Forest Regressor 100 Estimators · Non-linear	±2.74 AQI	0.952 (95.2% fit)	Deployed
Linear Regression Baseline Ordinary Least Squares	±14.55 AQI	Baseline	Reference

81.2%

Ensemble Random Forest reduces prediction error by 81.2% compared to linear regression baseline, capturing complex diurnal and multi-gas interactions.

Anomaly Detection Center loading…

Hybrid Isolation Forest & Z-Score

Isolation Forest: Isolates multivariate anomalies via tree depth.
Z-Score Explainability: Attribute the root-cause outlier sensor when threshold deviations exceed standard bounds.

Checking System…

Evaluating continuous telemetry parameters

Isolation Forest Anomaly Risk Score 0

Multi-Sensor Z-Score Explainability

Pollution Source Diagnostic Match —

Assessing Covariance…

Analyzing multi-sensor covariance, emission rate of change, and diurnal cycles…

Multi-Sensor Covariance Attribution Heuristic Confidence

Vehicular Exhaust 72%

Biomass Combustion 18%

Marine / Background 10%

Environmental Telemetry

System Menu