Why Every Data Scientist Should Learn Django in 2026

#!/usr/bin/env bash
set -euo pipefail
python -m venv .venv && source .venv/bin/activate
pip install django scikit-learn joblib
django-admin startproject churnsite .
python manage.py startapp predictor
# add 'predictor' to INSTALLED_APPS in churnsite/settings.py
python manage.py migrate
python manage.py createsuperuser
python manage.py runserver

# train_model.py: creates model.joblib for the Django app
import joblib
import numpy as np
from sklearn.linear_model import LogisticRegression

rng = np.random.default_rng(0)
X = rng.uniform([0, 100], [36, 2000], size=(500, 2))   # tenure, spend
y = (X[:, 0] < 6).astype(int)                          # toy churn rule

model = LogisticRegression().fit(X, y)
joblib.dump(model, "predictor/model.joblib")
print("saved predictor/model.joblib")

from django import forms

class ChurnForm(forms.Form):
    tenure_months = forms.FloatField(min_value=0, max_value=120)
    monthly_spend = forms.FloatField(min_value=0, max_value=100000)

from pathlib import Path

import joblib
from django.http import JsonResponse
from django.shortcuts import render

from .forms import ChurnForm
from .models import Prediction

MODEL = joblib.load(Path(__file__).with_name("model.joblib"))   # loaded once

def predict(request):
    form = ChurnForm(request.POST or None)
    result = None
    if request.method == "POST" and form.is_valid():
        d = form.cleaned_data
        risk = float(MODEL.predict_proba([[d["tenure_months"], d["monthly_spend"]]])[0][1])
        Prediction.objects.create(**d, churn_risk=risk)
        result = round(risk, 3)
    return render(request, "predictor/predict.html", {"form": form, "result": result})

def predict_api(request):
    form = ChurnForm(request.GET)
    if not form.is_valid():
        return JsonResponse({"errors": form.errors}, status=400)
    d = form.cleaned_data
    risk = float(MODEL.predict_proba([[d["tenure_months"], d["monthly_spend"]]])[0][1])
    return JsonResponse({"churn_risk": round(risk, 3)})

# predictor/models.py
from django.db import models

class Prediction(models.Model):
    tenure_months = models.FloatField()
    monthly_spend = models.FloatField()
    churn_risk = models.FloatField()
    created_at = models.DateTimeField(auto_now_add=True)

    def __str__(self):
        return f"{self.churn_risk:.2f} at {self.created_at:%Y-%m-%d}"

# predictor/admin.py
from django.contrib import admin
from .models import Prediction

@admin.register(Prediction)
class PredictionAdmin(admin.ModelAdmin):
    list_display = ("created_at", "tenure_months", "monthly_spend", "churn_risk")
    list_filter = ("created_at",)

# Run inside: python manage.py shell
from django.db.models import Avg, Count
from predictor.models import Prediction

print(Prediction.objects.count(), "predictions saved")
print(Prediction.objects.aggregate(avg_risk=Avg("churn_risk")))
high = Prediction.objects.filter(churn_risk__gt=0.5)
print(high.count(), "high-risk predictions")

# Export to pandas for deeper analysis
import pandas as pd
df = pd.DataFrame(Prediction.objects.values())
print(df.describe())

Project: Churn risk web app with Django

- Train a scikit-learn model; save with joblib
- Django form with validated inputs
- View that loads the model once and predicts
- Prediction table via the ORM; admin with filters
- JSON endpoint for other apps
- Login so only staff can use it
- Notebook analysing saved predictions
- README with screenshots and setup steps

Week 1  Django basics: project, app, views, URLs, templates
Week 2  Forms and the model view; load model once
Week 3  ORM model, migrations, admin, login
Week 4  JSON endpoint, tests, analysis notebook, README

# predictor/tests.py   Run: python manage.py test
from django.test import TestCase

class PredictApiTests(TestCase):
    def test_valid_input_returns_risk(self):
        r = self.client.get("/api/predict", {"tenure_months": 2, "monthly_spend": 500})
        self.assertEqual(r.status_code, 200)
        self.assertIn("churn_risk", r.json())

    def test_invalid_input_rejected(self):
        r = self.client.get("/api/predict", {"tenure_months": -1, "monthly_spend": 500})
        self.assertEqual(r.status_code, 400)