A Practical-ish Introduction to Data Science
summary
Argues that data science needs theory before tinkering and demonstrates it with real tools and numbers: uses scikit-learn and pandas inside a Jupyter notebook, imputes missing ages by honorific median, trains a decision tree on Titanic features (age, sex, class), holds out 25% for testing, sees initial 75% accuracy, prunes the tree and raises accuracy to 75% to 80% (5 point gain), highlights overfitting, shows simple algorithms and feature engineering actually produce usable results.
AI-written summary · about
chapters (24)
- 0:00Introduction
- 0:42Who am I
- 1:40What is Data Science
- 9:40Machine Learning Algorithms
- 17:05Decision Trees
- 19:58Kmeans clustering
- 21:13Sum of squared errors
- 22:18Clustering
- 23:00Kmeans
- 23:24Key takeaways
- 24:32Practical example of data science
- 24:48Why Titanic
- 25:29Hypothesis
- 26:10Python
- 26:42Jupiter Notebook
- 27:30Pandas
- 27:56Statistics
- 28:50Feature Engineering
- 32:18Data reflection
- 33:15Gender
- 33:54Age
- 35:08Building a model
- 35:47Machine Learning
- 38:22Takeaways
related talks
45mA Practical Introduction to Data Science
47mA Practical(ish) Introduction to Data Science
44mA Practical-ish Introduction to Data Science
40mEveryone can Data! - A Practical Data Science Intro
42mI'm Starting in DataScience – Where to Begin?
28mThe Coding Portion of Data Science
27mUsing Machine Learning to solve a classification problem with scikit-learn
19mData science demystified