ISSUE 455 · October 3, 2023

Introducing Data Elixir Pro 🙌🏼

It's live! Data Elixir Pro gives you a GPT-4 assistant that’s powered by the Data Elixir Archives. It's knowledgeable on all topics from the newsletter over the past 4 years and will help you with coding, analysis, visualization, troubleshooting, writing, and more. Plans start at just $12/month. Details here >>

Posts & Tutorials

Data-driven news discourse analysis with Python

In this two-part tutorial, Karlis Kanders shows how to track and analyze news and public discussions about specified topics. Part 1 covers data access and how to perform simple topic analysis. Part 2 explores ways to analyze the discourse in finer detail. This is easy to follow and the techniques could be applied to any text data and topic of your interest.
Data science at Nesta | Karlis Kanders

24 hours in an invisible epidemic

Part story-telling and part quantified-self, this post is an awesome look at the data gathered by the American Time Use Survey. The survey asks tens of thousands of Americans to track one day of their lives and it provides a lot of detail about what those days look like. Also, see the followup, "Behind the scenes with 24 hours in an invisible epidemic."
The Pudding | Alvin Chang

The Most Unexpectedly Good and Bad TV Episodes

Nice tutorial showing how to access and analyze IMDB data using R. There isn't anything fancy here but it's very well done and could be easily extended to explore other aspects of the IMDB data. Includes lots of code and screenshots along the way.
JLaw’s R Blog

Sponsored Link

Writing 100s of lines of code to gather external API data? 

Stop building and managing API integrations and data pipelines and return to shipping features. Access any data source, store, or ML model from a single endpoint — SQL, GraphQL, or REST.

How often does Roy Kent say "F*CK"?

Great post (and screencast!), showing how to use Poisson regression and bootstrap resampling to find confidence intervals for when Roy Kent (from Ted Lasso) drops F-Bombs. The screencast is a step by step code-through while Julia talks about her decisions along the way. This is a fun post and the ideas here are particularly useful with small datasets.
Julia Silge

SQLite insert speedups, from the useful to the absurd

How fast can you insert data into SQLite? Faster than you probably think. This post explores ways to get the best insert performance out of SQLite generally; what's in your control, and what isn't. There's a linked index at the top or skip to the end for the most important tips.
Voidstar

Fine, I'll run a regression analysis

Of course, data should win an argument, right? Well, not on the Internet. In this post, Nate Silver describes his approach for making arguments with data and how sometimes he ends up more in the weeds than he'd like. Ultimately, he shows here that COVID vaccination rates are strongly predictive of COVID death rates.
Nate Silver

Build an API in the cloud in less than 200 lines of code

Cloud tools and Python packages have become so powerful that you can build a scalable, cloud-based API in less than 200 lines of code. In this post, you’ll see how to use Google Cloud, Terraform, and FastAPI to deploy a queryable data API on the cloud.
Arthur Turrell

Webinar: Hot takes and heartbreaks from business and data leaders

Join Benn Stancil, Field Chief Technology Officer of ThoughtSpot and co-founder of Mode on October 10 at 9 a.m. PT for a live panel with leaders from Matillion and Jasper.ai. They will get candid about the common misconceptions and stereotypes data leaders face and how to collaborate more effectively with your business leaders. RSVP here.
// sponsored

Tools & Code

Vizro 

Vizro is an open-source Python toolkit that's designed to be a simple, low-code way to create scalable visualization applications. It's low-code, but even so, it offers a lot of flexibility for advanced solutions.
GitHub | McKinsey

Outlier

WiFi without internet on a Southwest flight

Great discussion about WiFi, hacking, and how to access data that you probably shouldn't. The post by James Vaughan that kicked off the discussion is also worthwile.
Hacker News

How did you like this issue of Data Elixir?

👎    1     2     3     4     5   👍

Was this email forwarded to you?  Sign up here >>