Wind Turbine Power Production

[Project assessment or Machine Learning and Statistics Module - GMIT 2020]

This repository contains a web service that uses a machine learning approach to make accurate predictions of wind turbine power from given wind speed based on the data set powerproduction.

The project has the following features:

• Jupyter notebook with three models trained on the data set powerproduction.
• Python script running a web service based on the best performing model.
• Dockerfile that builds and runs the web service in a container.

Submitted by: Olga Rozhdestvina (Student No: G00387844)

Lecturer: Ian McLoughlin

Programming Language used: Python

Set up

---

Applications used for completion of the project are The Jupyter Notebook, Visual Studio Code, cmder

Distribution of the Python used is Anaconda Python distribution.

How to run the code

---

1. Make sure that you have Python installed
2. Download or clone current repository "Machine-Learning-and-Statistics-Project"
3. Open Command Interpreter and get into correct directory
4. Install packages and run the app:

   • If using a virtual environment:

     Linux

     pip install -r requirements.txt
     export FLASK_APP=flask_server
     python3 -m flask run

     Windows

     pip install -r requirements.txt
     set FLASK_APP=flask_server
     python -m flask run

   • If using Docker:

     docker build . -t wind-power
     docker run -d -p 5000:5000 wind-power

     In case of an error, change the first line of the Docker file for your version of Python.

5. To view the model analysis run Jupyter Notebook and open Power_production_models.ipynb.

License

---

This project is licensed under the MIT License - see the LICENSE.md file for details