🚀 Building an End-to-End Iris Classification Pipeline using Airflow, Flask, and Docker

In this blog, we’ll walk through how to build and deploy a machine learning pipeline using the Iris dataset, orchestrated with Apache Airflow, served with a Flask API, and packaged with Docker.

📁 Project Folder Structure

airflow-final/
├── Airflow_Setup_Steps.txt
├── Dockerfile
├── app.py
├── iris-train-model.py
├── iris_pipeline_dag.py
├── requirements.txt
├── iris-airflow-main.zip
├── data/
│   ├── iris_train.csv
│   └── iris_test.csv
├── model/
│   └── iris_model.joblib

⚙️ Step-by-Step Breakdown

✅ 1. Data – data/iris_train.csv

This is the standard Iris dataset used to train the model.

✅ 2. Training Script – iris-train-model.py

import pandas as pd
from sklearn.ensemble import RandomForestClassifier
import joblib

data = pd.read_csv('data/iris_train.csv')
X = data.drop(columns=['species'])
y = data['species']

model = RandomForestClassifier()
model.fit(X, y)

joblib.dump(model, 'model/iris_model.joblib')

✅ 3. Flask API – app.py

from flask import Flask, request, jsonify
import joblib
import numpy as np

app = Flask(__name__)
model = joblib.load("model/iris_model.joblib")

@app.route('/predict', methods=['POST'])
def predict():
    data = request.get_json(force=True)
    prediction = model.predict([np.array(data['features'])])
    return jsonify({'prediction': prediction.tolist()})

✅ 4. Airflow DAG – iris_pipeline_dag.py

from airflow import DAG
from airflow.operators.bash_operator import BashOperator
from datetime import datetime

dag = DAG('iris_training_pipeline', start_date=datetime(2023, 1, 1), schedule_interval='@daily')

train_model = BashOperator(
    task_id='train_model',
    bash_command='python3 /opt/airflow/iris-train-model.py',
    dag=dag,
)

✅ 5. Dockerfile

FROM python:3.9

WORKDIR /app
COPY . .

RUN pip install -r requirements.txt
CMD ["python3", "app.py"]

✅ 6. Requirements – requirements.txt

flask
scikit-learn
joblib
pandas

✅ 7. Airflow Setup – Airflow_Setup_Steps.txt

Contains steps to setup Airflow, place the DAG, and run the scheduler and webserver.

🧪 Testing the Flask API

docker build -t iris-api .
docker run -p 5000:5000 iris-api

✅ Conclusion

This project delivers a reproducible and scalable pipeline that can easily be extended to larger problems. Future steps could include integration with S3, model versioning via MLflow, and alerting for drift detection.