Sesame-Production

Introduction

The data used in this workflow has been obtained from the United Nations Food and Agriculture Organisation public data repository.

The objective of this ML Workflow is to accurately predict the Production quantity of sesame seeds provided the Area harvested and the Yield Expected.

This steps of this workflow consists of

• Extracting the data from a remote storage.

• Feature extraction, preprocessing and transformation.

• A machine learning algorithm is applied to the transformed data and saved in a serialized format.

• Finally the last step of our pipeline is model evaluation.

---

The metrics used for this project are

• Coefficent of Determination (r²)
• Root Mean Squared Error (rmse)

---

How to build this workflow

• First you need to clone this repository using:

git clone https://github.com/ahmed-gaal/sesame-production.git

• Then you need to create a virtual environment using:

python3 -m venv env

• Activate your virtual environment using:

source env/bin/activate

• After activate your virtual environment, install the project dependencies using:

pip install -r requirements.txt

• Add the original data to your environment variables:

export DATA='https://drive.google.com/uc?id=1pC5Md6KKJYCkSb32BV3xV-SkEDg2icff'

• Finally, create necessary changes in either train script or feature extraction script and reproduce the workflow using the following command:

dvc repro

To launch the interactive web application, click ⇢ here

---