LETTERBOXD COLLECTION WHEEL
(A Claude Code and Processing Experiment)
A radial visualization where each movie is a slice • Distance from center shows rating

Retro Pixels, AI Insights: A Personal Media Exploration

Matthew White - 7/24/2025

There's something beautifully paradoxical about my relationship with media. By day, I design digital interfaces and solve complex UX problems with code, pixels, and algorithms—often collaborating with AI tools like Claude Code to rapidly prototype and iterate on solutions. By night, I walk over to my shelf and browse through my collection of 410 physical movies, the same way I used to wander the aisles at Blockbuster.

This parallel became clear to me recently as I was organizing my physical media collection and decided to visualize it using Processing—the same creative coding framework I first discovered in 2006 when I was trying to make art with algorithms. Back then, I had ambitious ideas for data visualizations that my coding skills couldn't quite execute. Years of trying, practicing, and learning with other languages and frameworks taught me to think creatively about problems and communicate my visual intent. Now, with AI tools like Claude Code, I could finally leverage that creative strategy and design thinking to bring those half-formed concepts to life. Sure, I could use D3 or other modern frameworks, but returning to Processing felt right—personal, even. Just like my collection itself.

Code as Canvas, 2006-2007

Back then, Processing felt revolutionary. Here was a way to bridge the gap between programming and visual expression, to treat code not as pure logic but as a creative medium. I remember those late nights in 2007, wrestling with for loops and sin() functions, trying to make something beautiful emerge from the mathematical precision of code.

Processing taught me that behind every digital interface lies a human decision about how information should feel, how data should move, how interaction should flow. Every curve drawn with bezier(), every color chosen with fill(), was a tiny act of interface design—long before I knew that's what I was doing.

That early experience shaped how I approach UX today. The same principles that made a Processing sketch compelling—visual hierarchy, rhythm, responsive feedback—are what make digital interfaces intuitive and delightful.

The Interface of Physical Media

But here's what I've realized: physical media has its own interface design language, one that's been refined over decades of human interaction.

Consider the simple act of browsing my movie collection. Each DVD case is a masterpiece of information architecture: the spine displays the title in carefully chosen typography, creating a visual rhythm across the shelf. The front cover serves as both artistic expression and functional UI—communicating genre, tone, and key information at a glance. The back cover unfolds like a detailed product page: synopsis, cast list, technical specifications, even screenshots that serve as visual previews.

This is interface design in its purest form. Every element has been considered for both aesthetic appeal and functional purpose. The physical dimensions of the case create natural constraints that force designers to prioritize information—much like responsive design principles force us to consider mobile-first thinking.

My collection spans from The African Queen (1951) to recent releases, with a heavy concentration in the 90s and 2000s—the golden age of DVD packaging design. During this era, studios were still learning how to translate the cinematic experience into a physical product you could hold, and the experimentation shows. Special editions with embossed covers, innovative case designs, elaborate booklets—each one a small study in user experience design.

The Ritual of Intention

There's a UX principle we often talk about called "choice paralysis"—when too many options overwhelm users and prevent decision-making. Netflix's endless scroll of algorithmic recommendations is a perfect example of this problem at scale.

But browsing my physical collection recreates the Blockbuster experience: a curated, finite set of options that I can actually process. I can see everything at once, spot patterns in my own taste, and make a decision without getting lost in an infinite feed of "Because you watched..." suggestions.

This constrained browsing creates what I call "productive friction." The extra steps—standing up, scanning titles, picking up a case—transform the paralyzing abundance of streaming into manageable choice architecture.

To be clear, I'm not a streaming luddite—I maintain a Netflix subscription, regularly go to theaters, and actively seek out new films and shows. But I often find that after scrolling through Netflix's recommendations, I've already watched everything that truly appeals to me and end up returning to my physical collection anyway. The curated constraint of my shelf becomes a refuge from the paradox of choice that infinite catalogs create.

It's the same principle behind good form design: sometimes we want users to slow down, to really think about their choices. The most important actions shouldn't always be the easiest ones.

Data as Artifact

When I visualized my collection in Processing—the same tool that first taught me to think about data as something visual and experiential—I wasn't actively searching for new patterns. Instead, the visualization became a mirror, helping me discover and understand the types of films I genuinely enjoy. It wasn't about finding hidden insights, but about seeing my personal taste laid out in a way that made intuitive sense to me.

Each disc or VHS tape in my collection is both content and container, story and object. Unlike digital files that exist as abstract data, these movies have weight, texture, presence. They occupy physical space in my home, creating a living archive of aesthetic choices and emotional connections.

But perhaps more importantly, DVDs and Blu-rays offer something streaming services have largely abandoned: curated context. Director commentaries, behind-the-scenes documentaries, deleted scenes, alternate endings—these bonus features create a contained universe around each film. When I want to dive deeper into a movie's world, I'm not pulled into the endless rabbit hole of YouTube theories or Wikipedia articles. Instead, I get the filmmaker's intended supplementary content, designed specifically to enhance that single viewing experience.

This is intentional design at its finest. The disc creates a boundary around the experience, letting me go deep without distraction. There's no autoplay pulling me toward the next thing, no algorithm suggesting what I should watch next. Just the film and its carefully curated companion materials, allowing for focused exploration without the cognitive overhead of the entire internet.

Interface Lessons from Analog

Working with physical media has made me a better digital designer. Here's what DVDs taught me about interface design:

Organization as experience design

Alphabetical shelves prioritize findability—like a search-optimized interface. Genre groupings create discovery zones where Blade Runner sitting next to Alien might spark a sci-fi mood I didn't know I was in. Director-based organization tells stories about artistic evolution. Each arrangement strategy creates a fundamentally different user experience with the same content. Digital products that lock users into a single organizational metaphor miss this insight.

Visual rhythm creates cognitive patterns

The way DVD spines create a typographic landscape across my shelf teaches me about scannable hierarchies in ways no design book could. Thick cases break the rhythm like bold headers break up text blocks—both guide the eye and create natural pause points.

Empty space signals intention—or absence

The gaps in my collection work both ways: sometimes they represent conscious editorial choices, but other times they signal something missing. When I see the empty spot where The Matrix used to be before I lent it out, that gap feels broken, incomplete. Interface design has the same tension—users can sense when something is missing even if they can't articulate what. That feeling of incompleteness can be interpreted as "broken," "missing," or simply lost. Understanding when white space feels intentional versus when it feels like an error is crucial.

Context switching changes everything

Moving the same movie from the "comfort food" shelf to the "must revisit" shelf transforms its identity entirely. User-generated tags and categories in digital products should embrace this fluidity rather than forcing rigid taxonomies.

The Future of Physical Digital

As I continue building digital experiences by day and curating physical ones by evening, I'm struck by how these practices inform each other. AI and its ability to free designers to do amazing things excites me—I can curate prototypes that are very compelling now very quickly. But the ritualistic nature of "make something, look at what you made, and figure out how to improve it" still persists in any form, whether I'm arranging movies on a shelf or iterating on interface designs with AI assistance.

What remains constant across both practices is the power of empathy, memory, nostalgia, and experience. When I choose which movies deserve space on my shelf, I'm drawing on emotional connections and personal history. What makes a movie memorable goes beyond the scenes, sound, or genre—it's the feeling they evoke or the experience in which you first saw them. A movie can be objectively bad in execution, but if you selected it at the video store with your friends and went home to watch it together, that memory becomes as much part of the experience as the film itself.

When I design interfaces, I'm tapping into that same well of human understanding—anticipating how users will feel, remembering how I've navigated similar problems, creating experiences that resonate beyond pure functionality. Design and UX work the same way: the context, the emotions, the memories we create around interactions often matter more than the technical execution—although in UX, eventually, technical execution becomes critical. A broken interface kills the emotional connection no matter how thoughtfully designed the experience.

The future isn't about choosing between digital convenience and analog authenticity—it's about understanding what each does best and designing accordingly. I awlays say "design the best thing you can for the output you're using".

Some experiences benefit from frictionless efficiency. Others are enhanced by ritual, intention, and physical presence. The best interfaces—whether made of code or cardboard—understand their medium and respect their users' time, attention, and humanity.

Conclusion

My Processing visualization captures 150 of my 410 physical films, and an attempt to discover 150 moments of huamn-centered design spanning nearly a century. Each data point asks the fundamental question: How do we help humans connect meaningfully with the stories that matter to them?

The AI Collaboration Journey

Working with Claude Code revealed the delicate art of AI collaboration. At times, it hallucinated movie titles and struggled to capture the human essence of technological moments on screen. What surprised me most was how it would struggle with seemingly simple tasks after initial success, requiring multiple rephrases and recontextualizations. Whether this was a limitation of the AI or a result of my own imprecise prompting remained an open question—a reminder that AI collaboration is as much an art of communication as it is of technology. Creating authentic design challenges required endless iterations, transforming generic descriptions into nuanced observations that reflect how characters truly interact with technology.

Sometimes the answer is an algorithm, sometimes it's a shelf of carefully curated movies. The best collections—like the best designs—need both comfort and surprise, always leaving room for human intuition and unexpected discovery.


Above visualization created with Claude Code (2025) and Processing JS (2008). Dataset is my personal physical media collection from Letterboxd

See more at http://tchr01.github.io