Lab | Deploying a Data Science Project with Django

In this lab, you will create a web application using Django to deploy a machine learning model. You can use the Iris dataset classification model (same as used for Flask) or train and save another model of your choice.

Prerequisites

• Python 3.7+ installed
• Basic knowledge of Python and machine learning
• Virtual environment setup knowledge

Lab Instructions

Step 1: Project Setup

1. Create a new project directory and navigate to it
2. Create and activate a virtual environment
3. Install required packages: django, scikit-learn, pandas, numpy, joblib
4. Create a new Django project called ml_project
5. Create a Django app called predictor
6. Add your app to INSTALLED_APPS in settings.py

Step 2: Prepare Your Model

Choose one of the following options:

Option A: Use Iris Classification Model (Recommended)

• Use the same Iris dataset classification model from the Flask lesson

Option B: Build Your Own Model

1. Choose your own dataset and machine learning problem
2. Train a classification or regression model of your choice
3. Save your trained model as a .pkl file in your project root

Step 3: Create Views

1. Create views in predictor/views.py:
   • A home view to display the input form
   • A predict view to handle form submission and return predictions
2. Load your saved model in the views file
3. Handle user input validation and error cases

Step 4: Configure URLs

1. Create predictor/urls.py with URL patterns for your views
2. Include your app URLs in the main ml_project/urls.py

Step 5: Create Templates

1. Create a templates/predictor/ directory structure
2. Create the following HTML templates:
   • base.html - Base template with navigation and Bootstrap CSS
   • home.html - Form for user input (extends base template)
   • result.html - Display prediction results (extends base template)

Step 6: Test Your Application

1. Run Django migrations: python manage.py migrate
2. Start the development server: python manage.py runserver
3. Test your application in the browser
4. Verify that:
   • Input form displays correctly
   • Model predictions work as expected
   • Error handling works for invalid inputs
   • Results are displayed properly

Deliverables

Your completed Django application should include:

• Working web interface for model predictions
• Proper error handling for invalid inputs
• Clean, responsive design using Bootstrap
• Proper project structure with templates and static files
• Screenshot of your Django Web application running on Local Host.

Submission

• Upon completion, add your deliverables to git.
• Then commit git and push your branch to the remote.
• Make a pull request and paste the PR link in the submission field in the Student Portal.

Good luck!