ISSUE 443 · July 11, 2023
Talks & Conferences
Generative AI with LLMs: Hands-On Training feat.
This presentation by Jon Krohn is a great introduction to large language models. It's easy to follow and covers a lot of ground, starting with the technologies that got us to where we are today. The discussion continues with what modern LLMs are capable of, how to train and deploy LLMs, and ideas for getting commercial value from LLMs.
ODSC | Jon Krohn
Sponsored Link
Complete customer profiles in your data warehouse
RudderStack Profiles takes the SQL grunt work out of building customer profiles. You specify the customer traits, then Profiles runs the joins and computations for you to create complete profiles, so you can build better models, faster. And that’s just one use case. Get all the details.
Posts & Tutorials
How to Peek at Statistical Tests w/o Breaking Them
Continuously monitoring A/B tests results in a considerable over-estimation of statistical significance. Many tests that should not be considered significant get misinterpreted. Generally, p-values are deeply problematic and will only make you regret your decisions. This post explores the problems of peeking and ways to alleviate them.
Brandon Rohrer
Joins 13 Ways
Inner joins are common in the world of databases, but one thing about them is that everyone seems to have a different idea of what they are. In this post, Justin Jaffray explores different definitions of Inner Joins, ways to think about them, and ways to implement them. Some are arguably the same, but they’re all interesting nonetheless.
Justin Jaffray
Demystifying Text Data
Text data isn't structured in neat rows and columns like numerical data and it can be tricky to work with. In this tutorial, Saeed Esmaili walks through a few options, and in particular, shows how to use the python data transformation library called "unstructured."
Saeed Esmaili
CLI tools hidden in the Python standard library
Typical Python installations contain a lot of little tools that most people don't know about. This post shows how to find those hidden tools and discover what they do.
Simon Willison
Webinar: How to generate business intelligence leveraging Yelp's rich first-party data on AWS
Discover how to create actionable insights using Yelp's robust data sets to analyze your marketplace, your customers, and grow your business. Explore use cases on how businesses leverage this rich data with AWS Data Exchange to make strategic business decisions. Register now.
Date: July 19, 2023
Time: 10:00 AM (PST)
10:00 AM (BST)
10:00 AM (SGT)
// sponsored
Tools & Code
Makie
Makie is an open-source, modern plotting library for Julia. It's a general-purpose tool that's powerful and packed with features that make it easy to explore large datasets, create graphics for publications, build dashboards, create interactive visualizations for the web, and more.
Makie
Messy Dates
Most packages for working with dates in R expect the dates to be in the standard yyyy-mm-dd format. But dates are often messy. {messydates} is an R package that allows for date translations from unstructured text, such as “Second of April, two thousand and twenty”, “26 BC”, as well as inputs that have uncertain, approximate, and sets/ranges of dates.
messydates
Career
How to Do Great Work
If you collected lists of techniques for doing great work in a lot of different fields, what would the intersection look like? How do ambitious people decide what to work on? What do they avoid? What are their common stumbling blocks? How important is luck? Why do nerds have an advantage? Why do rule-breakers rule?... Great post!
Paul Graham
Resources
Unraveling Principal Component Analysis (PCA)
PCA is a popular technique for analyzing datasets that have lots of dimensions for each data point and is commonly used in machine learning. This new book starts with the basics and then dives deep into the fundamentals, including eigenvectors, the spectral theorem, singular value decompositions, low-rank decompositions and more.
Peter Bloem
For Python! An Introduction to Statistical Learning
The newest edition of this popular statistics text was released a few days ago and this time, it's tailored for Python users. Topics cover some of the most important modeling and prediction techniques and is intended to be accessible to a broad audience. Follow the links to download the PDF.
Gareth James, Daniela Witten, et al.
 |