Applied Python for Data Science and Analytics

Ditch the Fluff, Embrace the Challenge: Master Python for Real-World Data Science and Analytics

Created byJeff James

Last updated 5/2024

English

English [Auto],

What you'll learn

Moving beyond basic Python tutorials into solving complex and practical problems
Mastering both Python and Pandas, while being fluent enough to let AI help you solve problems
Parsing natural language, structured data and analyze any kind of data you may encounter in business or research
Develop confidence in problem solving and your ability to think through a problem
Using AI as a Copilot as we get into progressively complex problem solving

Course content

7 sections • 42 lectures • 9h 6m total length

Introduction - what's this course about? Should you take it?6:57
Applied Python for data science and analytics teaches hands-on coding, problem-driven learning, and AI-assisted approaches to build commercially viable data skills using real workflows and libraries.
1 - Sorting Customers , Getting our feet wet11:38
Make sure you have a google account and can access http://colab.research.google.com - I've provided links to starter and final notebooks. Let's get after it!
2 - Comprehensions, Summing, Conditionals12:22
3 - datetime module, adding complexity to customers, dictionaries11:30
Apply the datetime module to convert strings to date objects and sort purchases by date, then use a dictionary pattern to accumulate total spend per customer.
4 - defaultdict, intro to Pandas10:26
Master using a defaultdict to handle missing keys, then leverage pandas dataframes to group by species, sum purchases, convert strings to dates, and sort by date.
5 - More Pandas, Introduction to .map8:38
convert date strings to pandas datetime with to_datetime, sort by order date, and use map on a series to create a new datetime column.

Birthday Paradox Intro - Reading Data, Sampling12:22
https://colab.research.google.com/drive/15cKoyNt9U89yuU9rONnxSOy9p2WLENki?usp=sharing

Start here! Happy Birthday
Applied 1 - Birthday Paradox Solved12:20
Applied 1 - Birthday Paradox Variations15:00
Explore birthday paradox variations through Python examples, counting true and false outcomes without counters, using enumerate, default dict, and nested loops, then simulate birthday problems with range and random sampling.
Applied 1 - Birthday Outro and Gotcha!5:23
Explore the birthday problem using the random module to generate birthdays, test for duplicates, and debug by inspecting inputs, illustrating sampling without replacement.

Natural Language Intro and Tokenization12:49
Learning to wrangle text is a key skill - let's get started!
Token Frequencies Solved - Regexp Intro10:00
Let's get our feet wet with regular expressions, too.
Regexp Solved - Using AI to help Code12:36
https://colab.research.google.com/drive/1_7PJTyY6S-bFwmPEOPrxTAhkwOv_KcOJ?usp=sharing

Sometimes regular expressions can get tricky, especially if we're building them dynamically. Let's use AI to help us solve our problem.
%prun performance and Intro to Sentence Similarity Challenge13:51
Let's think about performance analysis and think about how we might compare each sentence to all other sentences.
Pairwise Sentences Partially Solved - itertools10:34
https://colab.research.google.com/drive/1_7PJTyY6S-bFwmPEOPrxTAhkwOv_KcOJ?usp=sharing

Rather than write 2 for loops, let's use itertools.combinations to help us iterate over all pairs of sentences and then compute the length of token intersections while storing them and creating a dataframe later on.
Pairwise Sentences Solved - Jaccard Challenge12:25
https://colab.research.google.com/drive/1_7PJTyY6S-bFwmPEOPrxTAhkwOv_KcOJ?usp=sharing
Alright, we did it! As a bonus, try to go from mere token interaction length to Jaccard similarity! https://en.wikipedia.org/wiki/Jaccard_index#Overview

Intro the Section, zip and bigrams + trigrams challenge13:16
Learn document similarity using out-of-the-box tools and the zip operator to build bigrams and trigrams from tokenized text, and prepare you for scikit-learn countvectorizer analysis.
Working with CountVectorizer and the Core idea behind document vector spaces13:14
Learn how count vectorization creates a bag-of-words document vector space with CountVectorizer in scikit-learn, building vocabularies, applying fit_transform to produce sparse matrices, and measuring document similarity with pairwise distances.
Working with numpy and building our own distance metric11:49
Dot Product, Cosine Similarity and Embeddings14:55
Learn how to compute cosine similarity between vectors using dot product and L2 norms, then derive cosine distance. See practical examples with NumPy arrays and embeddings.
Pairwise Distances, Numpy Array Indexing and .argmin/max14:04
Explore computing pairwise distances between documents with numpy, using a count vectorizer, comparing euclidean and cosine metrics, and using argmin to identify similar documents while handling self-similarity with fill diagonal.
Newsgroups Challenge Solved - More on Pandas14:59
Leverage the 20 newsgroups data from scikit-learn by converting the train dictionary to a pandas frame, vectorizing with CountVectorizer, and finding the five closest documents via Euclidean distances.

Parsing the Garmin dataset using beautiful soup13:56
Converting Points into DataFrame, reviewing comprehensions13:28
Pandas Type Conversions, Plotting the Workout14:58
Learn to convert a list comprehension to a dictionary, load GPS data into pandas with from_records, and perform type conversions and describe elevation and heart rate for exploratory workout analysis.
Computing Cumulative Time and analyzing correlations14:34
Pairplots using Seaborn and Using AI to help us compute cumulative distance12:45
Using Scipy to find Peaks and Troughs - continuing to code with AI14:48
Compute distances between GPS pings by converting lat-long into tuples and applying haversine. Analyze elevation with SciPy find_peaks for peaks and troughs, and adjust thresholds and units as needed.
Battling through to get the correct peaks and troughs indicated in our dataframe14:59
Identify peaks and troughs with SciPy, align their indices to the original dataframe, and map them to compute distances in feet while visualizing with plots.
Avg Lap Distance and Also computing total climbing using AI and numpy clip14:56
Final notebook https://colab.research.google.com/drive/1ZcO3JzfI59o1pY98ALqND3C5ifEyBs4y?usp=sharing - GREAT WORK

Challenges - Part 114:40
https://colab.research.google.com/drive/1dZmr6C9aNCsyj2JSGhc-YNRbYvjIpauH?usp=sharing
Challenges - Part 214:05
Learn efficient Python techniques for summing the first n numbers with range and a formula, and count color data from Advent of Code using parsing and regex.
Challenges Part 3 - Battling Data Parsing14:33
Learn how to parse messy data using regular expressions, dictionaries, and pandas, counting color occurrences across game rows and converting patterns into actionable data.
Final Solution Coded Up13:12

Intro to Supervised ML, Classification using XGBoost14:47
Explore supervised classification using XGBoost, decision trees, and random forests, with a focus on feature encoding, explainability, predictability, and production-ready modeling in a business context.
Training our XGBoost Model - Thinking about Decision Criteria -precision, recall14:53
XGBoost automated categorical handling document mentioned https://developer.nvidia.com/blog/categorical-features-in-xgboost-without-manual-encoding/

https://docs.google.com/spreadsheets/d/1RaoRSl-OH7N8gs6htBRXR5aqqP9TQ5RlOYAo7GMm4D8/edit#gid=1218296526 the document I'm using to describe business decisions and the link to true positives, precision, recall, etc...
More about Precision and Recall14:42
XGBoost automated categorical handling document mentioned https://developer.nvidia.com/blog/categorical-features-in-xgboost-without-manual-encoding/

https://docs.google.com/spreadsheets/d/1RaoRSl-OH7N8gs6htBRXR5aqqP9TQ5RlOYAo7GMm4D8/edit#gid=1218296526 the document I'm using to describe business decisions and the link to true positives, precision, recall, etc...
Final Precision Recall Discussion across Scenarios (and Kangaroos)14:13
Explore precision and recall trade-offs across scenarios from rare disease diagnosis to credit eligibility and self-driving stoplight detection, emphasizing recall-focused strategies and practical model evaluation.
Saving / Loading our Model and Making Single Predictions14:55
Save and load your trained model with pickle, deploy to an endpoint, and predict from a dict to a dataframe; consider unseen values like OTA partner.
Feature Importance, Dangers of Categorical Encoding w/unseen values11:51
Discover how XGBoost handles unseen categorical values and the risks of custom encoders, then interpret model decisions with feature importance and partial dependence analysis.
Analyzing and Interpreting Lead time and Avg Room Price per Night Using Pandas14:57
Explore interpretability by analyzing lead time and average price per room with pandas, group and bin data, and simulate outcomes to explain XGBoost feature importance for business decisions.
Analyzing Features using Simulation and Prediction (prelude to partial dep)13:58
Simulate perturbing lead time, predict on modified data, and count canceled versus not canceled to visualize the marginal impact via a partial dependence plot.

Requirements

Understanding the basics of Python
A Google account to use Colab Notebooks
Have some appreciation for problems data scientists and analyst must solve on a daily basis.

Description

Unlock the Power of Python for Real-World Data Science and Analytics

Are you ready to take your Python skills to the next level and tackle real-world data science and analytics challenges? Look no further than "Applied Python for Data Science and Analytics," a comprehensive Udemy course designed to bridge the gap between memorization and practical problem-solving.

In this course, you'll learn from Jeff James, a senior machine learning engineering manager with 15 years of applied data analytics and coding experience, who has also taught at the University of Denver. Andrew will guide you through the complexities of the Python standard library, Pandas, SciPY, and powerful machine learning libraries like scikit-learn, empowering you to solve open-ended problems with confidence.

Throughout the course, you'll dive deep into real-world scenarios, learning how to approach and solve challenges that go beyond the typical "table of contents" style video courses. You'll gain hands-on experience working with diverse datasets, applying advanced analytical techniques, and leveraging the full potential of Python's data science ecosystem.

Whether you're a data analyst, aspiring data scientist, or a developer looking to expand your skill set, this course will equip you with the tools and knowledge you need to excel in the field. You'll learn how to:

- Effectively utilize the Python standard library for data manipulation and analysis

- Harness the power of pandas for efficient data wrangling and exploration

- Apply statistical techniques using SciPY to gain deeper insights from your data

- Implement machine learning algorithms using scikit-learn to solve real-world problems

- Develop a problem-solving mindset to tackle open-ended challenges in data science and analytics

By the end of this course, you'll have a robust portfolio of projects showcasing your ability to apply Python to real-world data science and analytics problems. You'll be ready to take on complex challenges, drive data-driven decision-making, and make a tangible impact in your organization.

Don't miss this opportunity to learn from an experienced industry professional and elevate your Python skills to new heights. Enroll now in "Applied Python for Data Science and Analytics" and unlock your full potential in the world of data science and analytics!

Who this course is for:

Data Analysts and Data Scientists
Beginner Python Developers that want to develop deeper skills

Applied Python for Data Science and Analytics

What you'll learn

Explore related topics

Course content

Introduction and Getting Moving with Python6 lectures • 1hr 2min

Applied Problem # 1 - Birthday Paradox4 lectures • 45min

Natural Language Parsing and Analysis6 lectures • 1hr 12min

Natural Language Project - Using ScikitLearn6 lectures • 1hr 22min

Building Expertise with Pandas and Analyzing Data8 lectures • 1hr 54min

Some Additional Python Practice - Not required, But see if you can do it!4 lectures • 57min

Applied Machine Learning - WIP - But Come Learn!8 lectures • 1hr 54min

Requirements

Description

Who this course is for: