DatalytIQs Academy

Global Earthquake–Tsunami Risk Assessment Dataset

Written by

Collins Odhiambo

in

, , , , ,

Seismic Features, Temporal Trends, and Global Distribution Analysis (2001–2022)

Overview

The Global Earthquake–Tsunami Risk Assessment Dataset is a comprehensive, machine learning–ready dataset that records 782 significant earthquakes worldwide from 2001 to 2022.
Each record includes detailed seismic parameters such as magnitude, depth, intensity, and tsunami potential, making it a valuable resource for tsunami prediction, hazard assessment, and AI-driven geophysical modeling.

Developed under the DatalytIQs Academy Research Initiative, this analysis blends earth science, statistics, and data visualization to promote a data-driven understanding of natural disasters.

Dataset Highlights

Attribute Details
Period Covered 2001 – 2022
Total Records 782 earthquakes
Coverage Global (Latitude −61.85° to 71.63°, Longitude −179.97° to 179.66°)
Completeness 100% (no missing values)
Target Variable Tsunami indicator (0 = No, 1 = Yes)
File Format CSV (~41KB)

Tsunami Event Classification:

  • Non-Tsunami Events: 478 (61.1%)

  • Tsunami-Potential Events: 304 (38.9%)

  • Balanced Dataset: Ideal for binary classification and deep learning models.

1. Descriptive Statistics

Feature Mean Std Min Max Description
Magnitude 6.94 0.45 6.5 9.1 Earthquake strength (Richter)
Depth (km) 75.88 137.28 2.7 670.8 Focal depth
Significance (sig) 870.1 322.5 650 2910 Event hazard score
Latitude 3.54 27.30 −61.85 71.63 Geographic range
Longitude 52.61 117.90 −179.97 179.66 Epicentral coverage
CDI (Community Intensity) 4.33 3.17 0 9 Perceived shaking
MMI (Mercalli Intensity) 5.96 1.46 1 9 Structural impact
NST (Stations) 230.25 250.18 0 934 Seismic monitoring coverage
Year 2012.28 6.10 2001 2022 Temporal span
Tsunami (binary) 0.39 0.49 0 1 Target variable

Summary

The dataset shows an average magnitude near 7.0, capturing globally significant quakes.
A wide depth range (3–670 km) ensures both shallow and deep events are represented.
The balanced tsunami variable (39% positive) enhances its value for AI model training.

2. Magnitude Distribution

Interpretation

  • The histogram reveals a left-skewed distribution, showing that most global quakes fall between 6.5 and 7.2 magnitude.

  • Fewer events exceed magnitude 8.0, representing the rare mega-earthquakes (e.g., Sumatra 2004, Japan 2011).

  • The steep decline after magnitude 7.5 demonstrates the logarithmic nature of seismic energy release: every 1-point increase equals roughly 32× more energy.

This reinforces that while smaller quakes are frequent, large ones dominate damage and tsunami generation.

3. Magnitude vs Depth of Earthquakes

Insights

  • Shallow earthquakes (<100 km) dominate high magnitudes and are more destructive.

  • Deep events (>500 km) tend to have moderate magnitudes, indicating less surface impact.

  • The clustering confirms subduction zones as the main regions of high seismic energy release.

4. Earthquake Frequency Over Time

Observations

  • Annual events fluctuate between 25–55, with spikes in 2010–2015, aligning with several mega-quakes.

  • 2011 and 2013 recorded the highest global activity.

  • No linear trend is evident, emphasizing irregular tectonic release rather than time-based cycles.

Implication

Predictive earthquake modeling must therefore rely on real-time geophysical indicators (e.g., plate stress, GPS deformation) instead of purely temporal data.

5. Global Distribution of Earthquakes

Geographic Patterns

  • Most events cluster along major tectonic boundaries:

    • Pacific Ring of Fire — Japan, Indonesia, Chile, Alaska.

    • Himalayan–Eurasian Belt — India, Nepal, Tibet.

    • Mid-Atlantic Ridge — Oceanic spreading zones.

  • The color scale shows magnitude intensity — brighter points indicate stronger events.

This distribution visually confirms that tectonic boundaries are the Earth’s most active seismic regions.

6. Machine Learning and Policy Applications

  • Tsunami Classification Models: Predict tsunami potential using seismic features.

  • Hazard Mapping: Visualize global high-risk zones.

  • Predictive Analytics: Use AI to assess future seismic hazards.

  • Infrastructure Planning: Guide resilient construction policies in coastal nations.

  • Real-Time Alerts: Feed trained models into IoT sensor systems for early warning.

7. Data Quality & Scientific Value

  • Zero missing values across all 13 columns

  • Global spatial coverage (−180° to 180°)

  • Balanced tsunami cases (39%)

  • 28 major earthquakes (≥8.0 magnitude)

  • Suitable for ML, visualization, and disaster risk analytics

Acknowledgment

This analysis was conducted by DatalytIQs Academy, a multidisciplinary education platform specializing in Mathematics, Economics, and Geoscience Analytics.

Data Source: Kaggle — Global Earthquake–Tsunami Risk Assessment Dataset (2001–2022)
Tools Used: Python, Pandas, Matplotlib, Seaborn, and JupyterLab

“At DatalytIQs Academy, we transform seismic data into global foresight — empowering resilience through analytics.”
— Collins Odhiambo Owino, Founder

Comments

Leave a Reply

More posts