Dr. Albert Rapp

Thank you for sharing, Matt 🤗❤️

Can you provide an example of the y-axis being incorrectly scaled? I'm curious about what's wrong with ggplot's scaling 🤔

Nice work 🤗 and thanks for the shout-out. 🥳

Want to build AI Bots with R? Combining the {ellmer} package with the Shiny framework might be exactly what you need.

And I'm happy to share that I get to talk about how that works at this year's ShinyConf 🎉 If you want to join, you can still sign up at www.shinyconf.com/agenda

#rstats

Also I just noticed that my DMs were blocked 😅 should now be open in case you want to reach out ✌🏻 (unfortunately I can't DM you)

I do consulting work in that space ✌🏻

Always happy to help 🤗

3 Minutes Wednesdays Weekly bite-sized tips on DataViz, Shiny and Stats/Machine Learning.

All three steps are pretty cheap and easy to pull off. If you're curious, tomorrow my 3-minute newsletter will release a tutorial to guide you through every step of the way. You can still join 3300+ readers at buff.ly/KYAQYdv

Ever wanted to restrict your Quarto project to only a few select people? You can do so by

1️⃣ Hosting your project on Azure Static Web Apps
2️⃣ Creating your own user base via Azure's AAD B2C service
3️⃣ Configuring your app so that it only allows users from your user base

#rstats (1/2)

3 Minutes Wednesdays Weekly bite-sized tips on DataViz, Shiny and Stats/Machine Learning.

But I get how these steps can feel confusing if you never had to do something like this. That's why I've created a tutorial that shows you every step of the way.

It will be released tomorrow on my 3-minute newsletter. If you want to join 3300+ readers, you can sign up at buff.ly/rRTZXE5

3️⃣ Point the client's redirect URI to your web app on Azure
4️⃣ Place the client credentials into the app via environment variables

That's it. Once you've done that, Azure takes care of the rest and your project is hidden behind a layer of authentication. (2/3)

The nice thing about hosting a Quarto project on Azure is that you can easily enable social logins via vendors like GitHub, Apple or Google. All you have to do is

1️⃣ Add an "auth" section to your config file
2️⃣ Get a client from the identity provider (e.g. Google)

(1/3) #rstats #datascience

I used to think that securing a Quarto website is about writing lots of code. But really it only requires

✅ hosting on Azure static web apps,
✅ creating a JSON file inside of your Quarto project, and
✅ filling that file with values that you can look up in the docs.

#rstats #datascience (1/2)

❤️ thank you, Roland, for these kind words 🤗 honestly I think many people have social media fatigue and that's okay. I have it too. That's why I'm concentrating on making more entertaining videos instead of daily posts 🙂

Setting this up is pretty easy but a bit lenghty to describe here. That's why my newsletter will release a step-by-step tutorial on how to host Quarto projects on Azure tomorrow. If you want to join 3300+ readers, you can do so at https://buff.ly/3PRzat5

Hosting Quarto projects on Azure is an easy way to add user authentication to your project. That way, only selected users can access your web page.

All you need for that is
🔹 a GitHub repo,
🔹 an Azure subscription, and
🔹 Azure's static web apps service.

3 Minutes Wednesdays Weekly bite-sized tips on DataViz, Shiny and Stats/Machine Learning.

And I've created many tutorials about Step 2 in the last few weeks, so it's time to show you how Step 1 works.

Tomorrow, my 3-minute newsletter will teach you how to use AWS Textract. If you want to get bite-sized tutorials delivered to your inbox, you can join 3300+ learners at 3mw.albert-rapp.de

It's fascinating how easy text analysis has become in recent years. Nowadays, text analysis reduces to

1️⃣ feeding a document/image to an OCR service like Textract, and
2️⃣ sticking the results into an LLM.

No huge entry barriers for data scientists to dive into text processing.

#rstats #datascience

You might want to try {rhandsontable}: jrowen.github.io/rhandsontable/

The book will cover all of these things and today we're getting started with the basics. Like every week, there's also a video version of the chapter. You can check them out here:

Video tutorial: https://buff.ly/40Wel6a
Book chapter: https://buff.ly/42CXE0U

New chapter for the "AI With R" book 🥳 We're getting started with the {ellmer} package. This package has lots of features like

✅ tool calling to mix AI with traditional functions,
✅ structured output to extract infos, and
✅ async streams to build chat bots.

#rstats #AI #datascience

3 Minutes Wednesdays Weekly bite-sized tips on DataViz, Shiny and Stats/Machine Learning.

That way, you can unlock lots of unstructured data that was previously hidden in PDFs or Images.

If you want to see more details, tomorrow my 3-minute newsletter will give you a step-by-step guide with specific examples. If you're curious, you can join 3300+ readers at https://buff.ly/3PRzat5

But what if you have a PDF? In that case, you can use the pdf_convert() function from the {pdftools} package to make the PDF into an image. Then, you can use the previous steps.

That's it. Not really much you have to do.

3MW (Turn Text Into Structured Data With R & LLMs) Learn how to use R, {ellmer} and structured outputs instead of regular prompting to turn unstructured texts into structured data frames.

Then, you can enforce a structured output by defining new type objects.

Key functions here are type_array() and type_object().

If you've read my newsletter from last week, then you should have a good idea of how they work: https://buff.ly/4aBw6uJ

Let's imagine you have an image of an invoice. You can put that into one of {ellmer}'s content objects.

Key function here is content_image_file()

With AI, it's really easy to process invoices or other documents using R. Here's what you have to do:

#rstats #AI

Also, do you have any particular use case you'd like to see implemented in R? Let me know in the comments and I might just expand the materials a little 🥳

New Project: AI With R 🥳

I've been writing about how to use AI/LLMs with R. So I'm happy to share that I'm starting to collect all the materials in one central space. You can read the first two chapters at https://buff.ly/3PQNzWz

#rstats #AI

Thank you for spreading the word about my tutorial ❤️

First Mr beast video I'd watch 🤣