Cheat Sheet for Python, Machine Learning, and Data Science

December 21, 2020 Sunil Ghimire

Over the past months, I have been gathering all the cheat sheets for Python, Machine Learning, and Data Science. I share them from time to time with teachers, friends, and colleagues, and recently I have been getting asked a lot by some of the followers on Instagram (@_tech_tutor & @aihub_), so I have managed and planned to share the entire cheat sheet collection. To give context and make things more interesting, I have added descriptions/excerpts for each major topic.

Note: These cheat sheets have recently been re-designed into Super High-Resolution PDF. These cheat sheets are very helpful as a quick reference for the concepts.

1. PYTHON CHEAT SHEET

Python is a most popular general-purpose, high-level programming language which was created in 1991 by Guido van Rossum and developed by Python Software Foundation to concentrate on the readability of code with its extensive use of white space. The Python development team was inspired by the British comedy group Monty Python to make a programming language that was fun to use. Python was named after the BBC show “Monty Python’s Flying Circus” as he was a big fan of the TV show.

Used for various purposes such as developing, scripting, generation, and software testing, this language is mostly used in the field of Artificial intelligence as it is open-source, free, dynamic, portable, easy to code, and integrated language (python can be integrated with another language like c, c++, etc.) Python, being a cross-platform language, can run on various platforms like Linux, Windows, Mac, Raspberry pi, etc. Python is much efficient as developers are allowed to write programs with fewer lines of code than other programming languages.

Python code is executed line by line at a time so it is also known as an interpreted language. Python relies on indentation, using white space, to define the scope, whereas Other programming languages often use curly brackets for this purpose. We can split the program into various modules which can later be used in other programs.

2. DATA IMPORTING

It is expected that data scientists will develop high-performance machine learning models, so bringing or importing the data to a Python environment is the starting point. The data scientist can only clean, visualize, wrangle, and build predictive models only after importing the data.

In this cheat sheet, you will learn the tips and techniques to import data like CSV Files, Text Files, Excel Data, Data from URL, and SQL Database into Python.

3. DATA CLEANING

The key purpose of Data Cleaning is to detect and remove errors and redundant data to create a consistent dataset. This increases the consistency of the analytic training data and makes for accurate decision-making.

Honestly saying, Data Cleaning is a time-consuming operation and the most data scientist spends a great amount of time for improving the quality of data.

In this cheat sheet, you will learn various techniques and tricks to identify and classify for data cleansing.

i. Missing Data
ii. Irregular Data (Outliers)
iii. Unnecessary Data – Duplicates, Repetitive Data, and more
iv. Inconsistent Data – Addresses, Capitalization, and more

4. JUPYTER NOTEBOOK

Jupyter Notebook is a web-based interface that helps you to build and exchange documents containing live code, visualizations, informative text, and equations. Also, Jupyter Notebook includes statistical modeling, numerical simulation, machine learning, data cleaning, and transformation.

In this cheat sheet you will learn the following techniques:

i. Demonstrate the scientific model to someone
ii. Experiment with the models
iii. capture and visualize the scientific journey
iv. share the scientific concept
v. operate the complex ideas
vi. make an impressive product demo
vii. teach to use the feature/product
viii. engage in training & learning
ix. communicate a certain message

5. FUNDAMENTAL PYTHON LIBRARIES

The Python Library is a collection of features and techniques that helps to perform several activities without writing code. Today, more than 137,000 python libraries exist. Python libraries play a critical role in the development of computer learning, data processing, data visualization, data, and image manipulation applications. If you are a beginner in the Python domain, I suggest you to learn how to use python for data analytics and data visualization.

Some of the popular Python Libraries are listed below:

a. NumPy

NumPy is an open-source library in Python that refers to Numerical Python that aids in numerical and calculations and computations, mathematical, data science programming, engineering, and scientific. The most popular for conducting mathematical and scientific operations is NumPy. For multi-dimensional arrays and multiplication of matrices, NumPy works perfectly well.

b. Pandas

Pandas is a software library that is used for data analysis, manipulation, and cleaning. For various kinds of data, the Python Pandas library is suited well. Such as:

In this cheat sheet you will learn the following techniques:

i. Tabular data with heterogeneously-typed columns
ii. Ordered and unordered time series data
iii. Arbitrary matrix data with row & column labels
iv. Unlabelled data
v. Any other form of observational or statistical data sets

Note: The container that a Pandas data object sits on top of a NumPy array

c. SciPy

SciPy is a Python library built on the NumPy extension to solve scientific and mathematical problems. Also helps the user with a wide range of high command levels to control and visualize data because SciPy provides easy and convenient manipulation of N-dimensional arrays.

Note:

i. SciPy consists of all numerical code whereas NumPy array consists of array data and basic operations.
ii. SciPy provides new data science features whereas NumPy library is useful for simple calculations in data science.

6. DATA VISUALIZATION

Visualization of data provides an insight into data. As a famous saying “A picture is worth a thousand words”. For a data scientist, data visualization is an incredibly significant and mandatory phase.

Knowledge of the following libraries is the perfect way to proceed with Data Visualization.

a. Matplotlib

Matplotlib is a Python library for creating 2D plots and 3D array plots. Although Matplotlib library is written in pure Python. Although Matplotlib Library is written in pure Python, to provide good performance even for large arrays, it makes extensive use of NumPy and other extension code. We can build scatter plots, line plots, bar graphs, histograms, stack layouts, pie charts, and 3D-image processing plots in Python by using Matplotlib.

A handy cheat sheet of Matplotlib with Python, including all the Matplotlib knowledge required for data visualization

b. Bokeh

Bokeh is an interactive visualization Python Library that provides high-performance web browser charts and plots.

Benefits of Bokeh:

i. Bokeh enables you to rapidly and efficiently create complicated statistical plots by simple commands.
ii. Bokeh gives you output in different mediums, such as a server, notebook, and HTML.
iii. Visualization written in other libraries such as seaborn, ggplot, and Matplotlib can be converted by Bokeh.
iv. Bokeh has the versatility to apply visualization layouts, interactions, and different styling options.
iv. Bokeh visualization can also be integrated into the flask and Django app.

A cheat sheet of handy Bokeh with Python, including all the Bokeh knowledge required for data visualization

c. Plotly

Another Python Visualization Library known as plotly is an open-source plot library that supports more than 40 different types of plots covering a wide variety of mathematical, financial, scientific, geographic, and 3-dimensional use-cases.

A handy cheat sheet of Plotly with Python, including all the Plotly knowledge required for data visualization

7. BIG DATA

Big Data is a blend of organized, half-organized, and unorganized data gathered by organizations, which can be utilized for data, predictive analysis, and advanced computational applications in the field of machine learning. They are additionally described by 6Vs: and a short description of them is given below.

i. Volume – Volume characterizes the amount of large data. The size of data can range from very low to high.
ii. Variety – Variety characterizes information kinds of enormous information, which incorporates organized and unstructured information, for example, content, sound, video, sensor data, etc.
iii. Velocity – The velocity accentuates the constant handling intensity of huge information for big business needs.
iv. Veracity – It refers to the prerequisite of the right type of information as it is depended upon for all further investigation.
v. Variability – Inconstancy Data can be in a similar structure yet having diverse semantics. Visualization – Data ought to be easy to process and decipher to infer knowledge out of it.

a. PySpark – RDDs

A basic data structure of Spark is Resilient Distributed Datasets (RDD) which is an immutable set of distributed objects. In RDD, each dataset is split into logical partitions, which can be computed on separate cluster nodes.

Properties of RDDs are:

i. Immutable (Unchangeable Objects)
ii. Partitioned (Logical Division of data)
iii. Fault-Tolerant (Lineage – a record of all the improvements that need to be made to the RDD, including where the data needs to be read from)
iv. Generated by operations of Coarse-Grained (operations applied to all elements in the dataset)
v. Lazily analyzed (RDDs can only be determined if any activity is called)
vi. Can be Persisted (Users will specify which RDDs, i.e. memory or disk, they can reuse and select a storage strategy for them.)

b. PySpark – SQL

In Spark, PySpark SQL is a module that supports the programming language of Python and also handles structured data. PySpark offers APIs that help heterogeneous data sources to read Spark Framework data for processing. It can be extended to a very large volume dataset and is extremely scalable.

Features of PySpark SQL

i. Speed
ii. Powerful Caching
iii. Real-Time
iv. Deployment
v. Polyglot

8. MACHINE LEARNING

Artificial intelligence can be interpreted as adding human intelligence to a machine. Artificial intelligence is not a system but a discipline that focuses on making machines smart enough to tackle the problem as the human brain does. The ability to learn, understand, images are the qualities that are naturally found in Humans. Developing a system that has the same or better level of these qualities artificially is termed as Artificial Intelligence.

Machine Learning is a subset of AI. That is, all machine learning counts as AI, but not all AI counts as machine learning. Machine learning refers to the system that can learn by itself. Machine learning is the study of computer algorithms that comprises algorithms and statistical models that allow computer programs to automatically improve through experience.

“Machine learning is the tendency of machines to learn from data analysis and achieve Artificial Intelligence.”

Machine Learning is the science of getting computers to act by feeding them data and letting them learn a few tricks on their own without being explicitly programmed. Machine learning can be further classified into three types:

i. Supervised Learning
ii. Unsupervised learning
iii. Reinforced Learning

The key difference between Machine Learning and Artificial Intelligence

i. Artificial intelligence focuses on Success whereas Machine Learning focuses on Accuracy
ii. AI is not a system, but it can be implemented on the system to make the system intelligent. ML is a system that can extract knowledge from datasets
iii. AI is used in decision making whereas MLis used in learning from experience
iv. AI mimics human whereas MLdevelops a self-learning algorithm
v. AI leads to wisdom or intelligence whereas ML leads to knowledge or experience
vi. Machine Learning is one of the ways to achieve Artificial intelligence.

a. Sklearn

Sklearn or Scikit-learn is one of the popular Python library for machine learning which contains the most efficient tools for statistical modeling and machine learning including classification, regression, clustering, and dimensionality reduction.

9. DEEP LEARNING

Deep Learning is the subfield of Machine Learning which is used to process materials and data that would be either be expensive or impossible for the human team to process in a short amount of time.

Some of the popular Deep Learning Libraries are listed below:

a. TensorFlow

TensorFlow is an open-source library developed mainly for deep learning applications and also supports traditional machine learning by Google. TensorFlow was initially generated without having deep learning in mind for massive numerical computations. However, TensorFlow has proved to be very beneficial for the advancement of deep learning, so Google open-sourced it. TensorFlow programs work on building a computational graph and executing a computational graph.

b. PyTorch

PyTorch is the most popular Python library that enables the development of deep learning projects. PyTorch emphasizes simplicity and enables idiomatic Python to express deep learning models. Since PyTorch contains OOP, it makes it easy to keep track of and much more readable compared to TensorFlow.

c. Keras

Keras is an open-source library of neural networks written in Python which provides various versions of widely used building blocks of the neural network, such as layers, activation functions, objectives, optimization, and a host of tools to make it easy to deal with image and text data. Keras has support for neural networks that are convolutional and recurrent. Other common utility layers, such as dropout, batch normalization, and pooling are supported by Keras.

10. NATURAL LANGUAGE PROCESSING

Human beings are the most advanced PC’s on earth and our success as human beings is because of our ability to communicate and share information, that’s where the development of languages comes in. And talking about human languages is the most difficult language that exists. So, coming to the 21st century, data is generated in the form of text, images, video, and audio files on WhatsApp, Facebook, or other social media. And the majority of data exist in the form of text. Thus, the concept of NLP was introduced and it can be simply understood as a component of AI which is the ability of a computer program to understand human language as it is spoken.

☺ Thanks for your time ☺

What do you think of this “Cheat Sheet for Python, Machine Learning, and Data Science“? Let us know by leaving a comment below. (Appreciation, Suggestions, and Questions are highly appreciated).