Imagine turning your reading history into a treasure map. By feeding a list of your favorite books and movies to an AI assistant, you can uncover hidden patterns in what you love. From your subconscious attraction to unreliable narrators to your love for stories that begin at the end, you may be surprised by what an AI assistant can reveal.
Building a personal “taste atlas” helps you understand your reading self better. It can also surface blind spots in your cultural diet and point you toward unexplored literary territories you’re likely to love.
Why analyze your preferences? ⚡️
This isn’t just another recommendation engine. Netflix or Amazon may suggest what to watch or buy next based on viewing history, but a taste atlas goes much deeper.
It analyzes themes, narrative structures, and emotional resonance across media formats. It can reveal connections between novels you adore and foreign films you’ve never heard of, or help you articulate why certain stories stick with you while others don’t.
You can tune the atlas by adjusting the info and examples you give it. You can customize the analysis with your prompts, asking for particular kinds of observations or recommendations.
With AI’s help, you can map out your own universe of awesome. As you scout out gaps in your reading or movie watching, you can discover authors and films that expand your horizons.
Start by gathering your favorites 🤩
You need to provide an AI assistant with a list of at least 10-15 titles that resonate with you for meaningful insights; 30+ is better. Here are the fastest ways to gather them.
* Physical books or DVDs: snap a photo of your bookshelf. AI can read the titles. Or write a list of titles on paper. AI assistants can read handwriting surprisingly well.
* Digital readers: refer to your Kindle library, your “read” shelf on Goodreads, listen history on Audible, timeline on Libby, or any doc or spreadsheet you maintain with your favorites.
* Streaming: Apps like Likewise, Sofa, Listy, Listium, Letterboxd, Trakt, and Reelgood let you compile lists of favorites. You can use those collections to train your AI assistant.
* Use your voice: If talking jogs your memory, use conversation mode in ChatGPT, Claude, Google’s Gemini, or Microsoft’s CoPilot. Let the AI interview you about your favorite books or movies.
* Scan award lists. If you can’t think of favorites, check a list of Oscar-winning movies or book awards for reminders of what you’ve enjoyed.
Criteria: Consider titles you often revisit or recommend. Include recent favorites and older resonant ones. Give extra weight to those that provoked emotion, changed your perspective, or prompted action. Ideally, note not just the title but one or more aspects of a work that particularly resonated.
Prompt AI to analyze your list 🔎
Once you've compiled your list, use your preferred AI tool to uncover patterns in your literary tastes. Prompt the AI assistant for insights to advance your self-understanding. After that, ask it to help you discover more books/movies you'll love.
Start by writing a detailed prompt to elicit a thorough, subtle analysis of your taste in books or movies. Here’s an example you can adopt or adapt:
You are a perceptive literary critic and cultural analyst with deep knowledge of literature across genres and cultures. Carefully analyze the attached list of my favorite books for patterns. Think deeply about connections between titles and topics that might not be immediately apparent. Where you notice interesting patterns, explain your reasoning and cite specific examples.
Please analyze this list of my favorite books. Create a detailed literary taste profile that identifies:
Core Elements:
* Primary themes and topics
* Genre preferences and style patterns
* Narrative approaches and structure choices
* Character types and relationships
* Tone and emotional range
»»»» Upload a file with your list or paste it.
* Here’s a related prompt for film.
* Additional taste atlas prompts to enrich your analysis.
* Case study of a taste atlas I created for my book group.
Which AI tool to use? 🎯
* ChatGPT 4o worked well for me in importing Google Docs and PDFs with my favorites. Its analysis and recommendations were nuanced and helpful.
* Limitation: Occasionally, it suggested authors who were already in my existing lists, despite being prompted not to.
* Claude Pro provided an excellent overview of the kinds of books I’ve selected for the book group I facilitate over the past eight years. It helped identify gaps in our reading list and offered useful suggestions for future titles.
* Limitation: Some documents I tried to import, like my Readwise reading highlights, were too large to fit in a Claude Project I created for my taste atlas.
* Gemini 2.0 Experimental Advanced, Google’s newest model, was an excellent voice partner in analyzing my current reading interests.
* Limitation: 2.0 couldn’t yet import documents, but Gemini 1.5 could. It helpfully analyzed the Google Doc with my complete Readwise Highlights archive.
Use either free or premium AI tools for this analysis. For long book lists or extensive highlights, use a pro model for nuanced analysis.
Sponsored Message
Become the better version of yourself with GenYOU. Create AI-generated headshots, profile pics, or even full-body portraits right from your selfies. The advanced AI ensures your facial features remain accurate with every generation. Skip the photoshoot hassle and design personalized AI portraits your way.
Expand your taste horizons ✨
Once an AI tool has analyzed your book or movie preferences, prompt it to suggest new authors and titles. Ask about specific connections between the titles you liked and its recommendations, so you understand the rationale.
* Cultural leaps: Ask AI to identify authors who write like your favorites but in different languages or cultures.
* What’s missing? Try a prompt about negative space — what authors, titles, topics or genres are missing from your favorites. What notable titles might stretch your literary horizons?
* Bridges to the past: Prompt your AI assistant to suggest "bridge authors" who influenced the writers you enjoy. This is most effective if the authors on your list are well-known.
* Cross media: Ask for documentaries and feature films that share traits with your favorite books. To push the AI further, ask for plays and songs.
Next Steps 💫
Make it a project 💻
If you use ChatGPT Plus or Claude Pro, start a dedicated project (learn more) to house your taste atlas. This lets you refine and expand your analysis over time.
* You can also create a Perplexity Space to combine AI search with analysis.
* Or make a Custom GPT or an AI Poe bot to share a group taste atlas with a class, a book group, or others who share an interest.
* NotebookLM (read more) is another great tool for analyzing collections of your favorite works in AI-powered notebooks. It accepts files of up to 50,000 words, up to 200mb, so it’s especially useful if you run into file size caps on other platforms. It’s also uniquely able to generate an audio summary piece about your favorites. (The video accompanying this piece has an excerpt of an audio piece generated with my reading highlights).
Share for human insight 🧓
Share your taste profile with a friend or librarian. They’ll spot patterns the AI missed or suggest unexpected connections.
Expand to music and beyond 🎶
Once you’ve mapped your reading and movie preferences, try a similar approach for your favorite music, art, food, and other interests.
Case study: my own taste atlas 🗺️
How I gathered my favorites: I had a Google spreadsheet with 80+ books I appreciated, so I exported it as a PDF to import into AI tools. I’m also gradually exporting my Day One book journal. For my movie list, I used Listy [here’s why it’s useful], which let me export an image and a CSV file of 30 movies I liked. Some apps don’t enable list exports, which requires screenshotting or copy and pasting.
Behind the scenes: Here’s a ChatGPT analysis thread illustrating how I created an initial taste atlas with my book and movie lists. You’ll see that ChatGPT struggled to understand Listy’s CSV export, so I fed it a PDF instead. I’ve also experimented with Claude, NotebookLM and Gemini. Each offers unique insights, so it’s helpful to try more than one. My taste atlas has already been useful. My night table has a new pile of books inspired by connections to books and films I loved.
Leave a comment to share your own taste atlas or to add thoughts, questions, comments, suggestions, or helpful resources 👇
p.s. Want to try something new this year? I’m looking for curious collaborators to work on exciting new projects. I’m also hiring a paid part-time assistant. Answer two quick questions if you’d like to work with me in either capacity.
Get full access to Wonder Tools at wondertools.substack.com/subscribe