🎨 Can a machine help me hang my art?

Moving into a new apartment comes with plenty of challenges, but one of the more peculiar ones I've been wrestling with lately is where to place my wall art.

I've brought with me a collection of artwork—varying sizes, orientations, and styles—that I've curated over the years. The blank walls of my apartment feel like a canvas waiting to be adorned, but the thought of positioning everything “just right” is daunting.

The context

Hanging art is deceptively tricky. We've all seen examples of beautifully arranged walls, ones that appear so cohesive they feel like an extension of the room's personality. But there are also those that miss the mark—too cluttered, too sparse, or just plain awkward.

I can't help but wonder: what makes one arrangement more aesthetically pleasing than another? Is it purely subjective? Or is there some inherent logic—a set of unwritten rules—that we intuitively follow, whether we realise it or not?

I could dive into social media for inspiration, poring over photos of Pinterest-perfect walls. But even with reference images, translating someone else's vision into my own space feels like guesswork. Could technology help?

Idea

A while ago, I experimented with TensorFlow to identify number plates on cars during my commute. The model recognised the plates, extracted the characters with OCR, and used an API to fetch metadata from the DVLA. While entirely unrelated to wall art, it introduced me to the idea of leveraging AI for niche problems.

Could a similar approach help me solve my art placement dilemma?

1. Scrape a dataset: Start with thousands of images of aesthetically arranged wall art sourced from social media and interior design sites.
2. Train a model: Feed the images into a machine-learning model to identify patterns—relative sizes, orientations, spacing, and positioning of artwork.
3. Input my collection: Use this model to “advise” me. By inputting my specific pieces (their dimensions and orientations), the model could suggest an arrangement that aligns with the aesthetic patterns it's learned.

Why AI?

This idea fascinates me because these “rules” of art placement aren't easily defined. While there are guides (e.g., hang at eye level, use a grid system), they're often too rigid or fail to capture the subtle, biological appeal of a great arrangement.

Machine learning thrives in these grey areas. It doesn't need a formal set of instructions; it digests raw data and makes sense of patterns we can't always articulate.

So, in a way, an AI-driven layout tool might not just mimic human design intuition—it could distill it.

But is it human thought?

If this hypothetical model produced a layout for my walls, could I consider it “thoughtful”? Or is it simply regurgitating patterns it has been exposed to?

There's something poetic about using a machine, ultimately a human creation, to help me position pieces of human creativity on my walls. Maybe it's not about whether the model thinks. Instead, it's about amplifying my own thinking, giving me the confidence to know there's some logic behind the final arrangement.

Bottom line: I just want my art to look good on my walls—and maybe an AI can help.