Hharkheem/SPE-DSEAT-Africa-Datathon-2024-
Domain:
environment and energyRecord type:
project
Creator:
Hha
Host:
This project is a submission for the SPE Data Science and Engineering Analytics Technical Section (DSEAT) Africa Datathon 2024. The goal of the competition is to build a robust machine learning model that accurately predicts oil, gas, and water production based on historical well production data.
# SPE DSEAT Africa Datathon 2024 Submission
**Author**: Hakeem Salifu
**SPE Number**: 5554170
**Category**: Student
**Affiliation**: KNUST
**Course of Study**: Petroleum Engineering
**Date**: July 2024
---
## 🔍 Project Title
**Prediction of Oil, Gas & Water Production in the DSEATS Field Using Machine Learning**
---
## 📘 Overview
This project is a submission for the **SPE Data Science and Engineering Analytics Technical Section (DSEAT) Africa Datathon 2024**. The goal of the competition is to build a robust machine learning model that accurately predicts oil, gas, and water production based on historical well production data.
---
## 📊 Challenge Objective
The main objective of this datathon is to:
- Analyze historical production data from a synthetic oil field.
- Develop a machine learning model that can predict oil, gas, and water production.
- Evaluate the model based on prediction accuracy, innovation, and creativity.
---
## 🧠 Machine Learning Approach
The approach follows a complete ML workflow:
1. **Data Preprocessing**: Cleaning and preparing training and validation datasets.
2. **Exploratory Data Analysis (EDA)**: Understanding feature distributions and correlations.
3. **Feature Engineering**: Transforming and creating features to improve model performance.
4. **Model Development**: Training multiple regression models and tuning hyperparameters.
5. **Evaluation**: Using metrics like RMSE, MAE, and R² to assess performance.
6. **Prediction**: Generating forecasts for the validation dataset.
---
## 📁 Repository Contents
| File Name | Description |
|----------|-------------|
| `Hakeem_Salifu_2024_DSEATS_Datathon_5554170.ipynb` | Jupyter notebook containing full ML pipeline |
| `Hakeem_Salifu_2024_DSEATS_Datathon_5554170.csv` | Final predicted production results (oil, gas, water) |
---
## 📈 Model Evaluation Criteria
- **Effectiveness of ML techniques**
- **Prediction accuracy (RMSE, MAE, R²)**
- **Total production estimation (forecast + histor …