Data Science Interviews

1.48MB. 0 audio & 27 images. Updated 2020-09-29.

Description

This deck helps you study the material behind common data science interview questions. It focuses on data science theory, mostly statistics and machine learning, rather than practice (it contains no code). It contains general knowledge needed for problem solving, rather than specific problems. Updated 9/20: Revised to make notes simpler, more concise, and more accurate. Many more pictures. Cards now have a reference link for easy access to source material. Split large cards into several parts. Included:

Statistics (counting and probability, hypothesis testing, distributions)
Machine Learning (bias/variance, over/underfitting, regularization, bagging, boosting, gradient descent)
Supervised Learning (linear and logistic regression, naive bayes, decision trees, random forests, SVM, KNN, neural networks)
Clustering (k-means, DBSCAN, spectral, hierarchical, evaluation)
Metrics (precision, recall, F1, sensitivity, specificity, ROC, AUC, log loss)
Linear Algebra (linear independence, orthogonality, bases, eigenvectors, invertible matrix theorem, SVD)
Feature Engineering (feature selection, PCA)
Soft Skills / behavioral and experience questions

Questions are grouped into sub-decks by topic, such as stats, supervised learning, and clustering. You can study just the sub-topics you want to learn and skip those you already know. Questions are tagged as high-level (broad, important, conceptual knowledge), medium-level, and low-level (less common, details, equations). A good study strategy would be to start with all high-level questions, then move on to medium and low. Material is sourced from around the web and from Data Science Interviews Exposed by You et. al. As you are studying the material, one great list of actual questions (and answers) on which to test your knowledge is Data Science Interview Questions & Detailed Answers. Does not cover: Data wrangling; Programming, engineering; Databases, SQL; Natural Language Processing; Deep Learning; Recommender Systems; Bayesian Methods; Time Series Analysis; Anomaly Detection; Visualization; Calculus; or the very basics. Requires Anki 2.1+ for Mathjax equations.

Sample (from 117 notes)

Cards are customizable! When this deck is imported into the desktop program, cards will appear as the deck author has made them. If you'd like to customize what appears on the front and back of a card, you can do so by clicking the Edit button, and then clicking the Cards button.

Front	Explain bagging
Back	Bootstrap aggregatingTrain multiple models on subsamples and average predictions to reduce variance.Usually uses "strong," low-bias models.
Ref	https://en.wikipedia.org/wiki/Bootstrap_aggregating
Credit	https://en.wikipedia.org/wiki/Bootstrap_aggregating
Tags	high-level

Front	How does an artificial neuron (perceptron) work?
Back	Applies an activation function to the weighted sum of its inputs.\[ y = f \left( \textstyle \sum w_i x_i \right) \]Common activation functions are linear, step, sigmoid, tangent, rectified linear...
Ref	https://en.wikipedia.org/wiki/Perceptron
Credit	http://www.theprojectspot.com/tutorial-post/introduction-to-artificial-neural-networks-part-1/7
Tags	high-level

Front	How do decision trees work, high-level?
Back	Recursively split the data into groups based on most discriminating feature; each leaf gives a prediction.
Ref	https://victorzhou.com/blog/intro-to-random-forests/
Credit	https://victorzhou.com/blog/intro-to-random-forests/
Tags	high-level

After the file is downloaded, double-click on it to open it in the desktop program.

At this time, it is not possible to add shared decks directly to your AnkiWeb account - they need to be added from the desktop then synchronized to AnkiWeb.

Reviews

on 1608738031

Clear and concise questions and answers.

on 1600225780

Nice deck. Covers main concepts!

on 1586692853

Using it a lot! Thanks

on 1584546694

Good!

on 1581520587

Thank you

on 1562757741

Awesome

on 1550469199

Exactly what I was looking for!

on 1547303850

Thank you! Good job!

on 1535528578

Thanks a lot!

on 1524787200

quite a comprehensive deck, with some subtle questions