The Powered by Data Show

AI agents can already build dashboards, but they leave behind a pile of one-off code that data teams can't audit or maintain. Ben Jones talks with Dave Fowler (GM of Charts at Fivetran, founder of Chartio) and RJ Andrews (author of Info We Trust, Sr. Principal Data Graphics Designer & Engineer at dbt Charts) about dbt Charts, an open-source "markdown for dashboards" that lives in a charts folder next to your dbt models.

They cover:

• Why dbt Charts does for the charting layer what dbt did for transformation

• Why charts are a social technology, and why agents need the right "grooves"

• How 350 years of chart history, including a 1930s Soviet chart with red year labels, shaped smarter defaults for ordinary bar and line charts

• Ben's first dbt Charts dashboard, built with Claude Code on live USGS earthquake data

• Language-first BI, the "Chartifact," and themes that extend each other

• Listener questions on Streamlit and Databricks, and on how fast upstream model changes reach a chart

• Why it's free for teams of up to three, in the spirit of Tableau Public

Chapters

(0:00) Welcome: Dave Fowler and RJ Andrews

(1:05) What dbt Charts is, and why AI-built dashboards need it

(2:37) Doing for charts what dbt did for transformation

(3:10) A markdown for dashboards, in your dbt repo

(4:29) Charts are a social technology

(5:40) Giving agents the right grooves

(6:57) How the project started: Chartio, dbt, and Fivetran

(8:19) "Our CEO will only read a line graph"

(10:05) A 1930s Soviet chart and making charts work harder

(11:25) Using the axis as an encoding channel

(12:27) Label every bar: opinionated defaults you can override

(13:15) RJ's design corpus

(14:58) Demo: Ben's first dbt Charts dashboard

(16:18) Language first: every dashboard is YAML

(17:35) The design panel and the "Chartifact"

(18:46) Themes that extend each other

(19:55) Outliers that bust out of the chart

(20:43) Analog vs. digital design freedom

(21:59) Designing a system, not a chart

(22:51) Listener question: where dbt Charts fits next to Streamlit and Databricks

(24:21) An open standard and an open-core business

(25:16) Free tiers: lessons from Tableau Public, Jira, and Chartio

(27:41) Charts next to models: rename a column without breaking dashboards

(28:59) Listener question: how fast do upstream changes reach a chart?

(30:38) What's next, and why charts need other people

(32:25) Getting started at dbtcharts.com

(33:44) Custom charts and the Betty White chart

(34:33) Breaking the rules, and wrap-up

Links

Episode page and full transcript: https://poweredbydata.com/podcast/dbt-charts-better-data-graphics-with-ai

dbt Charts: https://dbtcharts.com

Getting Started tutorial: https://docs.dbtcharts.com/guides/getting-started/

Dave Fowler: https://www.linkedin.com/in/dave-fowler-chartio/

RJ Andrews: https://www.linkedin.com/in/raymondjandrews/

Claude Duet is Anthropic's platform for scheduling work you want Claude to do, currently in research preview. A week after getting off the waitlist, Ben Jones walks through his own account channel by channel:

• A daily brief that adds industry news, a quote, a brain teaser, and a "today in history" to the usual email and calendar heads-ups

• Weekly security and edge-case scans of the websites and apps he's built

• A coding tutor for daily Python lessons

• Flashcards built from his 11 courses: 243 of them, plus a list of typos it found along the way

• A refresh log that keeps four AI courses current

• Weekly newsletter research and podcast planning

• Month-end close with QuickBooks, and a draft of the quarterly Washington State excise tax CSV

His point: it isn't just reminders. The agent does the work with you, on a schedule. He also covers the limits. Duet runs on a schedule, not on triggers, and in its first week it occasionally missed or ran a task late.

Not sponsored by Anthropic. Recorded November 2026; features may change.

Chapters

(0:00) What Claude Duet is (research preview, waitlist, not sponsored)

(0:58) A tour of the interface: channels, groups, chat, and the task panel

(2:23) Customizing channels: names, instructions, icons, and colors

(3:01) The daily brief: email and calendar heads-ups

(3:35) Adding industry news, a quote, a brain teaser, and today in history

(4:41) Video ideas: the channel that helped plan this episode

(5:00) Website management: why what you build needs maintenance

(5:21) Weekly security scans of the Data Literacy site

(6:07) Scanning an inventory app for edge cases

(6:45) A coding tutor for daily Python lessons

(6:59) Flashcards for staying fresh on 11 courses

(7:37) 243 flashcards, and the typos they turned up

(8:06) A running refresh log for the AI courses

(9:23) Weekly research for the monthly newsletter

(9:56) Keeping the podcast on track

(11:19) Month-end close with QuickBooks

(11:45) Drafting the quarterly Washington State excise tax CSV

(12:13) Not just reminders: the agent does the work

(12:31) Scheduled, not triggered: how Duet checks for new email

(13:20) Rough edges: missed and late tasks in research preview

(14:09) Wrapping up, and a preview of the next episode

Links

Episode page and full transcript: https://poweredbydata.com/podcast/claude-on-a-schedule-my-first-week-with-claude-duet

Claude Duet (product page and waitlist): https://claude.com/product/claude-duet

Introducing Claude Duet: https://claude.com/resources/articles/claude-duet

Get started with Claude Duet: https://support.claude.com/en/articles/17631495-get-started-with-claude-duet

Note: this is a screen-share episode, so the video version on Spotify and YouTube shows the walkthrough.

Ben Jones turned his book Avoiding Data Pitfalls into a Claude connector, an MCP server with 78 rules that lets Claude check its own data work. The connector never sees your data. It gives the session the rules and review instructions, and Claude does the checking.

In this solo episode, Ben builds a dashboard with Claude's new dashboards artifact from U.S. Department of Energy vehicle registration data, runs it through the connector, and walks through what came back:

• A 2025 hybrid jump that looks like a classification change, not real growth

• A summary comparing against a baseline year from before a known break

• Big percent changes on tiny, rounded counts

• One color meaning three different things in adjacent charts

• Side-by-side charts with axes that don't line up

• Registration estimates described as vehicles on the road

• A missing color legend

…and the pitfalls the dashboard avoided.

He closes with two other ways to use the same rules: the free datapitfalls CLI on npm, and the Pitfall Detector in the Lab on dataliteracy.com.

Chapters

(0:00) Welcome, and why Avoiding Data Pitfalls needed a new form

(1:28) Data Pitfalls: a Claude connector so Claude can check its own work

(2:19) The data: U.S. Department of Energy vehicle registrations by fuel type

(2:42) Building a dashboard with Claude's dashboards artifact

(3:09) Adding the Data Pitfalls connector (and why it never sees your data)

(4:07) A tour of the dashboard Claude built

(5:06) Running the dashboard through all 78 rules

(5:34) Pitfall 1: the 2025 hybrid jump and measurement system drift

(6:55) Pitfall 2: a baseline year from before the break

(7:18) Pitfall 3: big percent changes on tiny, rounded counts

(9:08) Pitfall 4: one color, three meanings

(10:30) Pitfall 5: side-by-side charts with mismatched axes

(11:28) Pitfall 6: registrations aren't vehicles (the data-reality gap)

(12:22) Pitfall 7: a missing legend, and the pitfalls it avoided

(12:47) Applying all seven fixes

(13:11) From good to great: why fine-tuning decides adoption

(14:17) The datapitfalls CLI on npm

(14:47) The Lab on dataliteracy.com and the Pitfall Detector

(15:31) Demo: scanning a campaign bar chart

(16:22) Wrapping up, and guests coming soon

Links

Episode page and full transcript: https://poweredbydata.com/podcast/pitfall-proof-your-claude-dashboards

Data Pitfalls connector for Claude: https://claude.ai/directory/data-pitfalls

Pitfall Detector in the Lab: https://dataliteracy.com/lab/pitfall-detector

Data Pitfalls Detector: https://poweredbydata.com/tools/data-pitfalls

datapitfalls on npm: https://www.npmjs.com/package/datapitfalls

Avoiding Data Pitfalls: https://www.avoidingdatapitfalls.com/

Note: this is a screen-share episode, so the video version on Spotify and YouTube shows the demos.

After eight years on WordPress, we tore the whole thing down and rebuilt it ourselves. This is the full tour of the new Data Literacy Learning Academy.

We retired the WordPress site we'd run on WP Engine since 2018 and replaced it with a Next.js and Tailwind application built with Claude Code: not just a marketing site, but the learning management system itself. It wasn't a one-weekend build. It took months.

Already had an account on the old site? Your badges, completed lessons, and quiz scores all carry over. Just sign up again with the same email address, since passwords don't transfer.

WHAT'S NEW

• Every course refreshed, and now on a quarterly update cadence. A two-year-old AI course is ancient history, and a one-time rewrite doesn't fix that, but a standing schedule does.

• An AI Tutor scoped to the lesson on your screen, an AI Coach that reads your learner profile, and inline definitions for any term you highlight. All in beta, all grounded in every book and course we've written using RAG (retrieval-augmented generation).

• Our first fully free course: 21 Key Traits of Data Literacy. Rate yourself on both proficiency and importance, then get a learning roadmap that points at free resources, not just ours.

• More coming! Right now, we're playing around with a live SQL sandbox running right in the browser, built on SQLite. I used it to cross-check every query in a client course before we shipped it, something I could never have done on the old site.

• New pricing tiers, including a course-only option for people who'd rather skip the AI features entirely.

CHAPTERS

0:00 Welcome back: a summer spent building

0:34 Why we retired the WordPress site after seven years

1:00 The new stack: Next.js, Tailwind, and Claude Code

1:53 Accounts, badges, and carrying your progress over

2:17 Refreshing every course — and a quarterly update cadence

3:12 AI in the lessons: inline definitions and the AI tutor

4:20 Vectorizing the books and courses with RAG

4:48 The AI coach and personalized learning paths

5:22 A free course for everyone: 21 Key Traits of Data Literacy

6:16 Under the hood: Claude Code, GitHub, Vercel, and Supabase

6:56 Demo: an in-browser SQL sandbox for hands-on courses

9:35 New pricing tiers and what each one unlocks

10:26 Cutting WP Engine and HubSpot out of the stack

11:34 Coming next: the Avoiding Data Pitfalls open-source package

LINKS

The new Data Literacy Learning Academy → https://dataliteracy.com

Episode page, show notes, and full transcript → https://poweredbydata.com/podcast/introducing-the-new-data-literacy-learning-academy

Data Pitfalls Detector → https://poweredbydata.com/tools/data-pitfalls

datapitfalls on GitHub → https://github.com/bjonesdataliteracy/datapitfalls

Ben Jones on LinkedIn → https://www.linkedin.com/in/benrjones/

Team training and workshops → https://poweredbydata.com/contact

The Powered by Data Show is hosted by Ben Jones, founder of Data Literacy. New episodes on data, AI, and the humans in the middle of it all. Subscribe so you don't miss one.

#DataLiteracy #AILiteracy #DataAnalytics #ClaudeCode #NextJS #SQL

Every leader claims to be "data-driven," yet most organizations still can't turn data and AI ambition into operational reality. In this episode of the Powered by Data Show, Ben Jones sits down with Patrick McGarry, Federal Chief Data Officer at ServiceNow, to talk about his new book, The Adaptive Organization: Leading Change in the AI Era.Patrick makes the case that transformation rarely fails because the technology is wrong. It fails because organizations treat a human and systemic challenge as a technology problem. We get into the CATALOG Framework at the heart of his book, why governance is infrastructure rather than overhead, and why the real constraint on AI is the human architecture around it.Ben also presses Patrick on the harder questions: whether governance slows you down in a world where models change weekly, where the "AI frees people up for higher-value work" story breaks down, and what honest leadership looks like when the math on headcount doesn't add up.📘 Pre-order The Adaptive Organization (paperback, hardback, and ebook), available now on Amazon ahead of the June 8th launch: https://www.amazon.com/Adaptive-Organization-Leading-Change-Era/dp/1960907123/Check out Patrick's book site: https://mcgarrycdo.com/In this conversation:00:00 Meet Patrick McGarry and the "accidental" origin of the book03:14 The CATALOG Framework and why transformation is a socio-technical problem05:19 Governance as infrastructure, not overhead: the Deming argument06:52 Does governance slow you down? Addressing the skeptical executive08:39 Building two identities: experimenting fast and scaling responsibly09:33 Governance should scale to consequences, not calendars11:44 The elephant in the room: AI, headcount, and job displacement17:53 The single highest-leverage move a leader can make today19:35 Why adapting once isn't enough, and what "choose to lead" really meansAbout the show:Powered by Data is where Ben Jones hosts candid conversations with data leaders, practitioners, and innovators. Each episode digs into the real challenges and practical approaches behind data strategy, BI modernization, AI adoption, and building data-literate organizations, all at the intersection of humans, data, and AI.

The Powered by Data Show, Episode 003In this episode of The Powered by Data Show, Ben Jones sits down with Rachel Herrera of Hex to explore how AI is changing the day-to-day work of data analysts, and why that shift may be better news than many people think. Instead of spending all their time writing repetitive SQL, cleaning messy data, and rebuilding the same dashboards, analysts now have an opportunity to become curators of context, stewards of trust, and stronger storytellers for the business. Ben and Rachel discuss Hex’s vision for AI-first analytics, including Context Studio, Hex's new feature that empowers data teams to manage and continuously improve agent performance over time. They also dig into what analysts should be learning right now to stay relevant: business understanding, discovery skills, judgment, and the ability to communicate clearly with humans. If you’ve been worried about what the future of the data analyst's role, this conversation might just give you some hope!

The Powered by Data Show, Episode 002

Ellie Fields helped shape the self-service analytics era at Tableau. Now, as CEO & co-founder of Ridge AI, she’s helping build what comes next. In this episode, Ben Jones and Ellie dig into the big question of whether dashboards are dead, where conversational AI breaks down, and how dashboards plus data agents may work together to create a better analytics experience.

In this episode of The Powered by Data Show, Ben Jones and Alli Torban introduce the podcast’s new name and vision, exploring why the conversation around data has shifted from understanding to action. They discuss what’s holding BI back from its full potential, how AI is changing the way people interact with data, and what listeners can expect from the show going forward, including real-world stories from leaders navigating the intersection of humans, data, and AI.

Show notes: https://dataliteracy.com/pbd001/

In this episode of The Data Literacy Show, Ben Jones (CEO of Data Literacy) and Alli Torban (Senior Data Literacy Advocate) tackle a question they’re hearing everywhere:

“Is our data ready for AI — and what does ‘ready’ really mean?”

We introduce DF4AI (Data Foundations for AI), which is a practical 10-minute self-assessment that helps organizations understand their AI readiness without chasing perfection.

You’ll learn…

• Why you don’t need perfect data to start getting AI value

• The five key dimensions of readiness

• How weak foundations quietly derail AI initiatives

If your team is experimenting with AI (or feeling pressure to), this episode will help you move forward with clarity and avoid costly missteps.

Take the free DF4AI self-assessment: https://df4ai.com/

See full show notes here: https://dataliteracy.com/season-02-episode-02/

In this episode of The Data Literacy Show, Ben Jones and Alli Torban share 12 predictions for data literacy in 2026, based on key lessons from the past year. They explore what leaders should watch for as data and AI become more embedded in everyday work — from the growing importance of measurement and human skills to the evolving role of AI, storytelling, and organizational change.

Show notes: https://dataliteracy.com/season-02-episode-01/