This project focuses on analyzing geochemical data from volcanic rock samples to predict rock age using data analysis techniques and machine learning models. The dataset contains elemental compositions obtained from Electron Microprobe Analysis (EMPA) and Laser Ablation Spectroscopy across different laboratories.
Determining the age of volcanic rocks is a crucial yet complex task in geology. Traditional dating methods can be expensive, time-consuming, and destructive. There's a growing need for data-driven, non-destructive techniques that use chemical composition to estimate the rock's age accurately.
This project addresses the challenge by leveraging geochemical analysis and machine learning to develop a reliable predictive model for rock age estimation.
- ⛏️ Accelerates geological research by automating age prediction of rocks based on their chemical signatures.
- 💰 Reduces the cost and time associated with conventional radiometric dating techniques.
- 🌍 Assists mining and exploration industries in identifying promising rock formations quickly.
- 🧪 Enhances data-driven decision-making in academic and field-based geoscience applications.
- Data Cleaning & Preprocessing – Handling missing values, normalizing data, and structuring datasets.
- Exploratory Data Analysis (EDA) – Extracting meaningful insights using visualizations.
- SQL & Python for Data Processing – Structuring, querying, and analyzing data.
- Feature Engineering & Selection – Identifying the most relevant geochemical elements.
- Data Visualization – Creating insights through Tableau, Power BI, and Matplotlib.
- Model Evaluation & Interpretation – Analyzing model performance to derive actionable insights.
The dataset contains elemental compositions of volcanic rocks analyzed in different laboratories. The dataset is divided into:
- EMPA Elements: (NiO, F, CaO, SiO₂, Cr₂O₃, Na₂O, TiO₂, etc.)
- Laser Ablation Elements: (Mg#, Li, Be, B, Sc, Ti, V, Cr, Mn, Co, Ni, etc.)
- ✔ Identified and handled missing values in Laser Ablation data using KNN Imputer.
- ✔ Standardized features to ensure uniform data scaling.
- ✔ Used SQL to filter and preprocess rock element concentrations.
- ✔ Analyzed correlations to determine key elements impacting rock age.
This project focuses on analyzing geochemical data from volcanic rock samples to predict rock age using data analysis techniques and machine learning models. The dataset contains elemental compositions obtained from Electron Microprobe Analysis (EMPA) and Laser Ablation Spectroscopy across different laboratories.
- Data Cleaning & Preprocessing – Handling missing values, normalizing data, and structuring datasets.
- Exploratory Data Analysis (EDA) – Extracting meaningful insights using visualizations.
- SQL & Python for Data Processing – Structuring, querying, and analyzing data.
- Feature Engineering & Selection – Identifying the most relevant geochemical elements.
- Data Visualization – Creating insights through Tableau, Power BI, and Matplotlib.
- Model Evaluation & Interpretation – Analyzing model performance to derive actionable insights.
The dataset contains elemental compositions of volcanic rocks analyzed in different laboratories. The dataset is divided into:
- EMPA Elements: (NiO, F, CaO, SiO₂, Cr₂O₃, Na₂O, TiO₂, etc.)
- Laser Ablation Elements: (Mg#, Li, Be, B, Sc, Ti, V, Cr, Mn, Co, Ni, etc.)
- ✔ Identified and handled missing values in Laser Ablation data using KNN Imputer.
- ✔ Standardized features to ensure uniform data scaling.
- ✔ Used SQL to filter and preprocess rock element concentrations.
- ✔ Analyzed correlations to determine key elements impacting rock age.
The following plot represents the distribution of geochemical features in the dataset. This visualization helps in understanding data variability and detecting potential outliers.
This heatmap shows the correlation between different EMPA elements, allowing us to identify highly correlated features that may impact model performance.
Similar to EMPA, this correlation matrix highlights relationships between Laser Ablation elements. High correlations indicate dependencies that can be considered during feature selection.
This bar chart showcases the most important EMPA features contributing to the prediction model.
This plot highlights the top contributing features from the Laser Ablation dataset used in model training.
| Model | Accuracy |
|---|---|
| Random Forest | 92% |
| Logistic Regression | 78% |
| XGBoost | 95% |
| Deep Learning (CNN) | 97% |
- 🔹 Laser Ablation data provides deeper insights but introduces variability due to missing values.
- 🔹 Feature selection improves predictive accuracy, allowing better classification of rock age.
- 🔹 Data visualization helps in trend identification, which is crucial for geochemical studies.
- 🔹 Machine learning models can enhance geological research, improving efficiency in geochemical age classification.
- 🔹 Laser Ablation data provides deeper insights but introduces variability due to missing values.
- 🔹 Feature selection improves predictive accuracy, allowing better classification of rock age.
- 🔹 Data visualization helps in trend identification, which is crucial for geochemical studies.
- 🔹 Machine learning models can enhance geological research, improving efficiency in geochemical age classification.
📂 Predict_Rock_Age
┣ 📜 README.md # Project documentation
┣ 📂 data
┃ ┣ 📄 rock_geochemistry.csv # Processed dataset
┣ 📂 notebooks
┃ ┣ 📄 analysis_notebook.ipynb # Main Jupyter Notebook
┣ 📂 visualizations
┃ ┣ 📄 correlation_heatmap.png
┃ ┣ 📄 element_distribution_chart.png
┃ ┣ 📄 model_accuracy_chart.png
┣ 📂 models
┃ ┣ 📄 best_model.pkl # Saved ML model
## 📎 How to Run This Project
1️⃣ **Clone the repository:**
```bash
git clone https://github.com/yourusername/Predict_Rock_Age.git
pip install -r requirements.txt
jupyter notebook analysis_notebook.ipynb
##🔗 **Connect With Me**
💡 If you find this project useful, feel free to connect!
📩 Email: mhassan.abid2024@gmail.com
🔗 LinkedIn: [www.linkedin.com/in/m-hassan-abid]