knowledge-kitchen

Exam 2 Review - Database Design (No Slides, Please)

This document has been autogenerated from slides.

---

Database Design

1. Overview
2. MongoDB
3. Web Apps
4. Pandas
5. Conclusions

Overview

Format

The exam will be composed of two parts:

• A Google Form, similar to a Quiz

• A GitHub Classroom repository, similar to an assignment

Time

Start anytime and submit anytime within a 24 hour window.

Topics covered

The topics covered on the exam:

• Do you really need a slide about this?

Weighting

All topics will be weighted approximately equally in grading.

MongoDB

Key points

Main topics:

• an example of a NoSQL database

• easily imports CSV and JSON file formats

• document-oriented - records stored as Javascript objects written with JSON

• Javascript objects support nesting of values

• normal forms need not apply

CRUD

CRUD is still cruddy

• Create: db.collection.insertOne(document) or db.collection.insertMany(documents)

• Read: db.collection.find(criteria, projection)

• Update: db.collection.updateOne(criteria, changes) or db.collection.updateMany(criteria, changes),

• Delete: db.collection.deleteOne(criteria) or db.collection.deleteMany(criteria).

Basic statistics

As with all database systems, MongoDB can easily do some basic statistical operations:

• Count: db.collection.countDocuments(criteria)

The aggregation pipeline is used to calculate most other statistics…

Basic statistics (continued)

E.g., calculate average of all values in the salary_range_to field with no grouping:

db.collection.aggregate([
  {
    $group: {
      _id: null,
      avg_val: { $avg: "$salary_range_to" },
    },
  },
])

Same, but grouping by agency:

db.collection.aggregate([
  {
    $group: {
      _id: "$agency",
      avg_val: { $avg: "$salary_range_to" },
    },
  },
])

Aggregation pipeline

The aggregation pipeline allows for multi-stage transformations, where the output of each stage becomes the input for the following stage.

Stages may include operations such as:

• $match - to filter documents by criteria

• $count - to count the number documents at this stage

• $project - to calculate or select particular fields

• $group - to group documents by a common attribute

• $sort - sort the results by a given field(s)

Web Apps

Key points

Main topics:

• the web is the use of the HTTP protocol that allows a web browser to browse hypertext documents

• a client (e.g. web browser) makes requests (e.g. HTTP GET or POST requests), and a server (e.g. web server) issues responses to those requests (e.g. HTTP 200 OK).

• HTTP POST requests can contain a payload of body data, such as when the user submits a form. GET requests (the default type) cannot.

• flask and pymongo are useful Python modules when creating web apps.

HTML, CSS, Javascript

Web browsers can only “understand”/interpret 3 languages: HTML, CSS, Javascript.

• HTML is used to indicate the content and semantic meaning of the content of a given web page.

• CSS is used to indicate the styling of the content on a given web page.

• Javascript is used to indicate any interactive behaviors of a given web page (changes to make on user click, mouseover, keyPressed, etc).

All web servers must respond to incoming HTTP requests from a web browser with code in one or more of these types.

flask

flask is a Python web server module to help detect incoming requests and respond to them

• flask allows us to set up server-side routes - functions that run automatically when a browser requests a certain web address.

• flask routes give us easy programmatic access to information in the incoming HTTP requests, such as any form data

• flask makes it easy to generate templates: static HTML documents with data injected into them dynamically.

pymongo

pymongo is a Python module that helps the server code communicate with a MongoDB databases.

• allows for all typical CRUD operations to be handled on a MongoDB database using simple Python code.

• for example, when a web browser requests a web page, a flask-based server app can… fetch data from a database using pymongo and inject it into an HTML template that is placed into the HTTP response.

• another example, when a user fills in a form and clicks submit, the browser can make an HTTP POST request to a server, where a route in a flask app receives the form data, packages it up nicely into a document, and saves it into a MongoDB database.

Pandas

Key points

Main topics:

• powerful Python-based data analysis library

• stores data in Series and DataFrame data structures

• vector math!

• relies on numpy, wraps around matplotlib

• includes tools to help munge/cleanup the data.

• supports basic data analysis: statistical analysis including grouping.

Jupyter Notebooks

A convenient web-based IDE widely-used in the scientific Python community for writing and sharing code.

• code cells and Markdown cells

• ability to run cells in the web browser

• supported by many platforms, including GitHub, Visual Studio Code, JupyterLab, etc.

Conclusions

Thank you. Good luck.