This project analyzes and predicts earthquake activities in Washington State using historical data from 1904 to 2024. The dataset, sourced from the United States Geological Survey (USGS), provides a comprehensive record of seismic events, allowing for detailed analysis and predictive modeling of future earthquakes.
-
Identify Most Affected Cities:
- Analyze which cities in Washington State have been most impacted by earthquakes.
-
Plot Historical Earthquake Data with Fault Lines:
- Visualize earthquake data on a map of Washington State alongside known fault lines.
-
Statistical Analysis:
- Explore the frequency, magnitude distributions, and depth of earthquakes.
-
Predict Future Earthquake Probabilities:
- Use machine learning models to predict the likelihood of future earthquakes based on historical data.
- Data Description & Loading: Understanding the dataset and loading it into the environment.
- Initial Data Analysis: Explore basic statistics and identify key features of the data.
- Data Cleaning & Preparation: Handle missing values, outliers, and derive new features.
- Univariate & Multivariate Analysis: Explore relationships between variables and analyze temporal patterns.
- City Analysis: Identify cities with the highest and lowest frequencies of earthquakes.
- Time Analysis: Analyze patterns over time, including time of day and year trends.
- Location Visualization: Map earthquake epicenters and compare with fault lines.
- Geospatial Analysis: Use Moran’s I and other geostatistical techniques to identify spatial patterns.
- Model Preparation: Data preprocessing for regressive modeling.
- Various Models: Implement MLPRegressor, AdaBoostRegressor, and TensorFlow models.
- Ensemble Methods: Use ensemble techniques to improve predictions.
- Model Preparation: Data preprocessing for classification tasks.
- Various Models: Implement RandomForestClassifier, MLPClassifier, and various ensemble methods.
- Performance Evaluation: Assess model performance using accuracy, mean squared error, and classification reports.
- Earthquake Probability Mapping: Visualize predictions on a map of Washington State.
- Combined Layer Analysis: Integrate different layers of data for comprehensive insights.
- Summary of Findings: Discuss key insights and implications for future research and policy.
The project requires the following libraries:
- Data Handling:
pandas,numpy - Visualization:
matplotlib,seaborn,folium,geopandas - Machine Learning:
scikit-learn,tensorflow - Geospatial Analysis:
libpysal,esda,shapely - Miscellaneous:
warnings
To install all dependencies, run:
pip install pandas numpy matplotlib seaborn folium geopandas scikit-learn tensorflow libpysal esda shapely-
Load the Data: Place the dataset (
WA_Earthquake_data.csv) in the designated directory and run the notebook cells to load and preprocess the data. -
Run the Analysis: Follow the sequential steps in the notebook to perform EDA, statistical analysis, and machine learning.
-
Visualize the Results: Use the visualization sections to view earthquake patterns, prediction maps, and other insights.
-
Model Predictions: Experiment with different models and parameters to optimize predictions for future earthquake probabilities.
The dataset used in this project is sourced from the United States Geological Survey (USGS). It includes records of earthquakes in Washington State from 1904 to 2024, detailing attributes such as magnitude, depth, and location.
- United States Geological Survey (USGS) - Earthquake Data
- Scikit-learn Documentation - Scikit-learn
- TensorFlow Documentation - TensorFlow
This project is developed as part of a comprehensive analysis and predictive modeling study of seismic activities in Washington State. Special thanks to the USGS for providing access to the earthquake dataset.
This project is licensed under the MIT License.
For questions or feedback, feel free to reach out:
- Email: Adam.RivardWalter@gmail.com
- GitHub: adamw80