Reminder: My latest HOPPy (Hands-On Projects in Python) course, where you design and build your own game using agentic coding, starts on Sunday! I can't think of any better way to level up your skills, as well as have a project in your portfolio. Learn more at https://LernerPython.com/hoppy .
AI continues to dominate the news headlines. And with US midterm elections coming up, one topic has been in the headlines quite a bit, namely the number of data centers currently being planned and built. These data centers are often massive -- with at least one reported to be the size of Manhattan -- and local residents in many states have indicated that they don't want this sort of thing nearby.
The politics of data centers are crossing party lines, too, with Governor Josh Shapiro of Pennsylvania (a Democrat) and Governor Greg Abbott of Texas (a Republican) both backtracking from their previously enthusiastic endorsements of data-center construction.
Meanwhile, Donald Trump is all in on data centers, saying that anyone who doesn't want them is stupid and wants to be poor – a political argument that The Bulwark recently described as so tone deaf, "Next, we suspect he will launch into caps-lock tirades against ice cream, puppies, and the Beatles" (https://www.thebulwark.com/p/donald-trump-ai-hubris-is-a-blue-democratic-opportunity).
Meanwhile, Republican politicians are struggling to find a way to please both Trump and their constituents (https://www.nytimes.com/2026/09/29/us/republicans-trump-data-center-ai.html?unlocked_article_code=1.FFE.1X-0.AanfR5azyTjm&smid=url-share).
This week, we'll look at when and where data centers are being built. But there's a catch: We'll do it not writing the Pandas queries ourselves, but instead during it using agentic coding. (I'm partial to Claude Code, but you can use any agentic coding system you want, using any models you want.) The idea is that we'll pose questions to the agents and ask for analysis from them – with the output going into a Marimo (or Jupyter) notebook. Then we can look at their queries, and see if we can learn something.
This week's questions are deliberately harder than usual, because I'm assuming you'll ask agents to do the hard work for you.
Data and five questions
There are a number of data sources, none of them perfectly authoritative, that you (or your AI agent) might wish to use, including:
- Epoch: https://epoch.ai/data/ai-data-centers
- FracTracker: https://fractracker.org/2026/04/open-u-s-data-centers-tracker/
- Interconnection: https://www.interconnection.fyi/data-center
Paid subscribers, both to Bamboo Weekly and to my LernerPython+data membership program (https://LernerPython.com) get all of the questions and answers, as well as downloadable data files, downloadable versions of my notebooks, one-click access to my notebooks, and invitations to monthly office hours.
Learning goals for this week include: Using agentic coding to perform data analysis — and then understanding the techniques it used.
Here are my five questions and tasks for this week. I'll be back tomorrow with my solutions and explanations.
- Retrieve data from one or more of the above sources (or others, if you know of them), combining them into a single Pandas data frame. Keep only data about the United States. Using Plotly, create a bar plot showing the number of data centers in the United States each year.
- Using Plotly, create a bar plot showing (as of the most recent data) how many data centers there are per state. If you have data for it, use a stacked bar plot to break each state-count bar into pieces, showing how many centers are associated with each company.