COSC 2P08: Week 1 Notes

What is Data Science?

Interdisciplinary field used to extract knowledge or insights from data. The application of data centric, computational, and inferential thinking to understand the world & solve problems.

Is Data Science…

• Statistics?
  • Yes, use data to infer properties of the world
• Machine Learning?
  • Yes, use data to build algorithms that make predictions
• Computer Science?
  • Yes, use computational thinking and abstraction to manage complexity
• Science, Art, Engineering?
  • Yes, combines the scientific method, creative thinking, and the ability to solve complicated problems

Big Concepts in Data Science

• Data preparation and representation
• Efficient data processing
• Question formulation and experimental design
• Exploratory data analysis
• Modelling, parameter estimation, and statistical inference
• Various prediction methods: generalized linear models, decision trees, neural networks, clustering, PCA…
  • Overfitting, regularization, and cross validation

Principles Statistics in Data Science

• Experimental Design & Sampling
  • How do we collect data to accurately answer questions?
• Probability & Uncertainty
  • How do we quantify what we don’t know?
• Modelling
  • How do we distill the essential structure of complex phenomena?
• Inference & Prediction
  • How do we use the known to reason about the unknown?

Principles of Computer Science in Data Science

• Software Design & Debugging
  • How do we develop and maintain reliable & repeatable analysis?
• Abstraction & Algorithm Design
  • How do we break big problems into small problems?
• Computational Complexity
  • How do we tradeoff time and space to compute efficiently?
• Parallelism & Locality
  • How do we divide computation across resources?

Domain Knowledge

• What is the key questions/problems in the domain?
• What is the context of the data?
  • What data is already available?
  • How and why was it collected?
  • What is the schema and limitations of the data?
  • How can more data be collected/obtained?
• What is the underlying process that generates the data?
  • casual structure, dependencies, …

Data Scientists must be inquisitive and learn new domains quickly.

Data Science Lifecycle

High-level description of the data science workflow:

graph LR

id1[Obtain Data]
id2[Understand Data]
id3[Understand World]
id4[Ask Question]
id5[Reports & Data-products]

id1 --> id4
id2 --> id4
id2 --> id3
id2 --> id1
id3 --> id1
id3 --> id2
id3 --> id5
id3 --> id1
id2 --> id5

• Frames questions & design experiments
• Obtain and clean data
• Summarize and visualize data
• Inference and prediction

Info

End of lecture 1, the rest of the notes are taken from the slides.

What is Big Data?

• Big in rows (size n)
• Big in columns (dimensions p)
• Hard to extract value from
• Big data is high volume, high velocity and high variety information assets that demand cost-effective, innovative forms of information processing

Sources of Big Data

• Data as the by-product of other activities
  • Click trail, clicks before a purchase
  • Moving your body (Apple Watch)
  • Data were always being generated. They just weren’t being captured.
    • Cheaper, smaller sensors help
• Data as the primary goal of activities
  • Telescopes, Genetic sequences, 61 million person experiments
• Web
  • 20 billion webpages, each 20KB = 400TB
• Astronomy
  • Apache Point Telescope, 200GB/night
  • Large Synoptic Survey Telescope: 3 billion pixel camera
• Life Sciences
  • High throughput sequencer: 1TB/day

Implication for Statistics

• Little data, big data
  • Sampling still matters
• Everything is significant (The Starry Night)
• Inverting a matrix
  • (Stochastic) Gradient Descent (Ascent), BFGS, …
• Casual Inference

Implication for Computation

• Conventional Understanding of what is computationally tractable: Polynomial time algorithm (N^K)
• Now it is (N^k)/m, where m is the number of computers.
• For really big data: N*log(N)
  • Traversing a binary tree, sort, and search N*log(N)
  • Streaming application

Our Goals

• Prepare students for advanced Data Science courses in data-management, machine learning, and statistics (by providing the necessary foundation and context)
• Enable students to start careers as data scientists by providing experience in working with real data, tools, and techniques
• Empower students to apply computational and inferential thinking to tackle real world problems

Note

End of lecture 1.

Introduction to Python

What is Python?

• General purpose programming language
• Interpreted programming language
  • Code is translated and executed one line at a time
• Object-oriented language
  • Objects created from classes

Python 2 vs Python 3

Python 2 is a older version of Python.

Important

Python 3 is not backwards compatible. If you write a Python 3 program and run it using Python 2, there is a good chance it won’t work.

A Simple Python Program

# Print two messages
print("Welcome to Python!")
print("Python is fun.")

You can run this program using:

python file.py
# OR
python3 file.py

PIP - Package Manager

• Package-management software written in Python
• Used to install and manage package installations
• Pip connects to Python Package Index
  • Online repository of public Python Libraries
• We will be using pytest to test lab assignments, and you can install it using the following command

pip install -U pytest

Statement

A statement represents an action or a sequence of actions. The statement print("Welcome to Python!") is a statement to display “Welcome to Python!” in the console.

Indentation

The indentation matters in Python. Note that the following program would fail due to a IndentationError: unexpected indent.

  print("Welcome to Python!")
print("Python is fun.")

Comments

Note that a # Character starts a comment for the current line. 3 quotation marks ''' ''' will also open and close a paragraph comment (multi-line comment).

# This is a comment
 
print("Comment test") # This is also a comment
 
'''
This is a 
multi-line
comment
'''
 
This is not a comment

Special Symbols

Character	Name	Descriptions
()	Opening and closing parentheses	Used with functions, arrays, etc.
#	Hashtag/Pound sign	Starts a comment
" "	Opening and closing quotation marks	Enclosing a string
''' '''	Opening and closing quotation marks (3x)	Enclosing a paragraph comment

Proper Indentation and Spacing (for this course)

• Comments
  • Include your name, class section, instructor, date, and a brief description at the beginning of the program
• Indentation
  • 4 spaces
• Spacing
  • Use blank line to separate segments of the code

Programming Errors

• Syntax errors
  • Error is code construction
• Runtime error
  • Somewhere in the code causes the program to abort
• Logic error
  • Code produces incorrect result

Variables, Expressions, Assignment, & Simultaneous Assignment

# Compute the first area
radius = 1.0
radius = radius + 0
radius = r = 0
area = radius * radius * 3.14159
print("The area is ", area, " for  radius ", radius)
x, y = y, x # Swap x with y

Note

Wow, there are 119 slides covering the basics to Python. I already know the basics to Python. I don’t need to spend hours writing notes on this lol.