What Python is used for: web, data and AI

One language, many jobs

You have learned core Python. Companies use that same Python with extra packages for specific work. The examples in this lesson need those packages installed with pip, so they do not run in the browser here. Read them to see how familiar they look.

Web: Django and FastAPI

A web framework is a package that handles the repeated parts of a website: reading the address the user visited, talking to the database and sending back a page or data. Django is a large "batteries included" framework. FastAPI is a lighter framework for APIs. An API (application programming interface) is a web address that returns data, usually JSON, for apps to use.

# needs: pip install fastapi uvicorn  (does not run here)
from fastapi import FastAPI

app = FastAPI()

@app.get("/price/{item}")
def price(item: str):
    prices = {"chai": 15, "samosa": 20}
    return {"item": item, "price": prices.get(item, 0)}

# a Django view (inside a Django project; does not run here)
from django.http import JsonResponse

def hello(request):
    return JsonResponse({"message": "Namaste"})

Data: pandas

pandas is the most used package for working with tables of data. Its main object, the DataFrame, is a table with named columns, like an Excel sheet inside Python.

# needs: pip install pandas  (does not run here)
import pandas as pd

df = pd.read_csv("expenses.csv")
print(df.groupby("category")["amount"].sum())

Under the hood, groupby does what you can already write with a dict and a loop:

expenses = [("food", 250), ("travel", 120), ("food", 90)]
totals = {}
for category, amount in expenses:
    totals[category] = totals.get(category, 0) + amount
print(totals)

AI and machine learning

Machine learning means the computer learns patterns from examples instead of following rules you write by hand. scikit-learn is used for classic models (for example predicting house prices), and PyTorch for deep learning (images, speech, language). Many apps also call large language models through an API using a few lines of Python.

model.fit(examples, answers)    # learn from past data
model.predict(new_examples)     # guess answers for new data

Automation and scripting

Many people first use Python at work to automate boring jobs: renaming files, filling Excel sheets, sending reports, cleaning data. The standard library (os, pathlib, csv, json, datetime) covers much of this.