CSE 163

Memory

��Hunter Schafer

​

​

​

​

​

💬 Before class: What is your favorite breakfast?

🎵Music: Japanese Breakfast

Memory

Just a big array/list to store values

• Names of these regions are not important for us

2

Objects

• Data and objects are just chunks of memory
• When you construct an object, the computer gives you an appropriately sized chunk of memory
• The fields are just specific locations within a chunk to store data�
• This is why we need the difference between equality of value and equality of identity
• Equal value: memory chunks store the same values
• Equal identity: same chunk of memory�
• By creating objects, you are making your program use more memory

3

Memory Hierarchy

4

Disk Drive

5

Takeaways

• Understanding how memory works can help you write more efficient programs (or understand some strange bugs)
• Key takeaways
• Do follow principles of locality
• Avoid reading/writing to disk frequently
• Watch out for large datasets! Paging can slow down your program a TON.

6

CSE 163

Mid-Quarter Review

��Hunter Schafer

​

​

​

Overview

What is this class?

​

​

​

​

​

​

​

​

Competencies

​

1. More advanced programming concepts than in CSE 142 or CSE 160 including how to write bigger programs with multiple classes and modules.�
2. How to work with different types of data: tabular, text, images, geo-spatial, etc.�
3. Ecosystem of data science tools including Jupyter Notebook and various data science libraries including scikit-image, scikit-learn, and pandas data frames.�
4. Basic concepts related to code complexity, efficiency of different types of data structures, and memory management.

8

Quarter �So Far

Competencies: Program | Data | Data Sci. | Comp Sci.

​

• M0 - Introduction to Python
• Program
• M1 - Advanced Python + Data Structures + CSVs
• Program | Data
• M2 - Pandas
• Data | Data Sci.
• M3 - Data Science Libraries (Data Viz + ML)
• Data Sci.
• M4 - Classes, Objects, and Modules
• Program | Data
• M5 - Efficiency
• Comp. Sci.

9

0 - Intro to Python

Summary

Learn the basics how how to write Python programs

​

Topics: Program

• Hello world!
• Variables and Data Types
• Control Structures
• Strings (slices)
• Lists
• Documentation
• None

10

1 - Python cont. + Data Structures

​

​

Summary

Learn more advanced Python concepts including how to represent data in data structures. Process our first type of data called CSV.

​

Topics: Program | Data

• Files
• Advanced Lists
• Python Modes (REPL vs script)
• Jupyter Notebooks (Data Sci.)
• Tuple
• Set
• Dictionary
• CSVs
• List of Dictionaries format

11

[

{'name': 'Seattle', 'magnitude': 4},

{'name': 'Genovia', 'magnitude': 6},

{'name': 'Seattle', 'magnitude': 3.5}

]

​

2 - Pandas

Summary

Learn our first data science library, pandas, to help process CSVs.

​

Topics: Data | Data Sci.

• Pandas objects: DataFrame and Series
• Access Data
• Access column
• Filter with a mask
• .loc
• Aggregates (mean, sum, count)
• groupby
• apply
• Missing Data

12

3 - Data Science Libraries

Summary

Learn other common libraries for data science. Learned the principles of data visualization (seaborn) and machine learning (scikit-learn).

​

Topics: Data Sci.

​

13

• Make plots in seaborn
• Point of visualization
• Effectiveness of encodings
• ML vocab
• Features
• Model
• Learning Algorithm
• Training vs test set
• Classification vs Regression

​

• Categorical data
• Decision Trees
• Overfitting
• Hyperparameter tuning

4 - Classes, Objects, and Modules

Summary

A new way of organizing our code into objects with state and behavior.

​

Topics: Program | Data

• State (fields) and behavior (methods)
• References
• Define a class
• Construct an instance/object
• Value equality vs Identity equality
• “Special methods” (__init__, __eq__, etc.)
• Why the main-method pattern is necessary
• Inverted index
• TF-IDF

14

5 - Efficiency

Summary

How to understand the efficiency of our programs.

​

Topics: Comp. Sci.

• Limitations of Timing
• Big-O notation
• Complexity classes
• Profiling
• Memory management
• How computer memory works

15

CSE 163

Requests

��Hunter Schafer

​

​

​

Internet

Allows computers to talk to each other to communicate data

​

One way to use the internet is to use a browser (e.g. Google Chrome) to browse web-pages.

​

Everything is determined by a URL that identifies

• Which computers to talk to
• Which resource you want that computer to give you

​

17

https://courses.cs.washington.edu/courses/cse163/22wi/assets/images/pokemon.png

Resources

In most cases, a URL specifies either

• A file that we want to access
• Example: https://courses.cs.washington.edu/courses/cse163/20sp/files/hw6/images/puppy.png
• A special command to the particular service to return some meaningful information. This is called an Application programming interface (API).
• Example: http://api.open-notify.org/iss-now.json
• There are many types of APIs that return different things depending on what you ask, but it is a useful way for one computer to get data from another

​

18

API Demo

• requests module
• GET request
• Response codes
• Response data
• JSON data

​

For this example, we used the Open-Notify API

• Documentation: http://open-notify.org/Open-Notify-API/

19

HTML

A website is actually just text that’s in a very special format: HTML

​

Basic idea:

• Use “tags” around parts of the text to indicate meaning
• Example tags
• Paragraphs
• Images
• Links
• Etc.

​

​

20

HTML Tags

Generally, we use an open tag and a close tag to say indicate everything inside belongs to that type of data:

• Tags can have attributes (src, href)
• Some tags are “self-closing” like img

​

​

​

​

​

​

​

​

Wikipedia > View Source

​

21

<p>This is a paragraph.</p>

​

Here is an image: <img src=”dog.jpg”>

​

<p>

This is another paragraph but it has a

<a href=”http://www.puppies.com”>link</a>!

</p>

​