Claude is having a moment right now. The Super Bowl ad, pushing back on the Department of Defense, and generally reminding people that AI should help us think, not replace thinking.
I won’t drink all the Kool-Aid but if I have to choose between ChatGPT and Claude, I’m with Claude.
Here’s why.
Last fall, I joked that AI gave me a promotion to Super Senior Developer. I asked you: has AI given you a promotion? What new skills has it unlocked?
My hunch is that many of you can now be data analysts. The biggest barrier has always been the tooling — SQL, Python, Tableau. AI may have blown through this ceiling. To test this out, I created a sample nonprofit dataset with 500 people, 75 organizations, 6,000 payments, as well as events, volunteer hours, and memberships. The kind of data a real nonprofit might keep. You can use AI to explore the data and build a dashboard.
Try It For Yourself
Here’s how:
- Download the dataset (one file — includes everything)
- Open Claude or ChatGPT
- Upload the file and say: “Help me analyze this data.”
Then start asking questions:
- Which donors gave last year but not this year?
- What percentage of people who attend our events are also donors?
- Which months do we receive the most donations?
- What percentage of our revenue comes from donors vs grants vs services?
- What percentage of our donors are monthly givers vs one-time?
Play around. Get curious. See what you find.
The Catch
One reason you can analyze this data is because of its structure. People connect to payments. Payments connect to campaigns. Events connect to registrations. The tables are clearly defined and the relationships are clean.
If this were 15 random spreadsheets with inconsistent column names and duplicate records, no amount of AI would save you. AI needs thoughtful, structured data under the hood. If the dataset is wobbly, the analysis will be too.
(A good data model is built on good definitions.)
Claude vs ChatGPT
I tested this in both Claude and ChatGPT and Claude was far and away the winner. Anthropic has clearly invested heavily in data visualization — you get interactive charts directly in the chat, not the ugly matplotlib charts.
If you’re a nonprofit choosing between the two, my official recommendation is Claude. (And it’s not even close.)
After you’ve experimented, let me know what you find. What surprised you?
More importantly — what opportunities do you see? Do you feel like a data analyst yet?
Until next week,
Ted
