Skip to content

README.md

Latest commit

 

History

History
71 lines (57 loc) · 3.44 KB

README.md

File metadata and controls

71 lines (57 loc) · 3.44 KB

Web Application Development with Flask

In the modern day, most applications we interact with are web applications - client-server architectures delivered via HTTP and accessed with a web browser. As a data scientist we will generally not build such applications, but we do have to understand them and their components.

Learning Objectives

  • Distinguish between front-end, back-end, database, and what tasks are appropriate for which
  • Create a simple Python web application that exposes an API to URL endpoints

Before Lecture

Read up on Flask, a "microframework" for developing web applications. All this really means is that it's small and modular, and mostly just provides for URL routing and responses - for other things (templates, database, forms) you pick and choose your own other packages (and we'll give you some specific choices for this week).

This is in contrast with Django, the most popular "industry-grade" Python web application framework. Django is opinionated, and comes with built-in modules for pretty much everything.

There are more projects built with Django than Flask, but Flask is solidly second place in the Python web application ecosystem, and is a better minimal choice for a data scientist who just wants to get some results.

Live Lecture Task

We're going to kick off with an overview of the major parts of a web application (front-end, back-end, database), and then develop our own prototype web application using Flask and SQLite. A summary of the overall picture:

  • Front-end: the "look-and-feel" of an application, as well as any logic that is executed by client-side (browser) JavaScript (which has grown surprisingly powerful)
  • Back-end: the routing and "business logic", where responses (what the client sees) are built and APIs and databases are accessed
  • Database: the "source of truth", where data is persisted and updated

Assignment

  • Reproduce the steps from lecture: write and run a basic local Flask web application
  • Create a model for Tweet and User as demonstrated in lecture, and populate your local (SQLite) database with some invented data (at least 6 Tweets and 2 Users)
  • For tomorrow, make sure to sign up for the Twitter Developer API and Basilica, as we will use both APIs in our application to get more and real data!

Resources and Stretch Goals

  • Jinja2 is the dominant template engine, which you will use to build the look and layout of the pages in your application
  • Flask-SQLAlchemy is an ORM (object-relational mapping), and will let us both (a) not have to write SQL, and (b) use OOP to interact with data with multiple backends (SQLite locally, and PostgreSQL in the cloud)
  • SQLAlchemy Data Types are needed to build usable models that will translate to SQL
  • Making the front-end look nice is very much a stretch goal throughout this sprint - but if you are interested in it, Picnic CSS and Umbrella JS are nice lightweight and modern tools (alternatives to Bootstrap and JQuery respectively)
  • Experiment with SQLAlchemy models - you can add more fields of interest to the Tweet and User models and/or add different models for other purposes