Skip to content
Clark Hayashi

Hi, I’m Clark

I grew up in Honolulu, went to college in Seattle, and studied abroad in Japan and Italy.

  • Marketing
  • Analytics
  • Sports
  • Global Business
  • Program Management
AI-assisted ink and watercolor illustration of Tokyo, Diamond Head, and Seattle along one shoreline

My journey

Honolulu

Drag · select a landmark · zoom with + and −

Honolulu

Where I'm from

~2022

I grew up on Oahu, where community isn't a value statement; it's how things get done. Punahou basketball and a varsity co-captain year, a 199-hour Eagle project rebuilding a library, and a first job as a sales associate helping military families at the Navy Exchange.

Sunset over the Hawai'i coastline

The Hawai'i years →

Things I have made

All projects →
First-year game participation and next-year return rates, from 23 percent for 1-2 games to 84 percent for 11 or more A closer lookClose preview

Who comes back to play again?

Participation analysis from a player and manager

Seattle University Recreation

As a player and manager, I used my growing data skills to investigate problems in a program I cared about.

Games played in the first year

23% returned after playing 1-2 games in their first year

Association, not evidence that playing more caused a return

8 seasons · 4,169 player-years · 648 played in 2025

Explore the full project
A Zippy's Korean Fried Chicken plate lunch with rice and macaroni salad, the order the speculative campaign was built around. A closer lookClose preview

Could Zippy’s feel at home in Seattle?

Zippy’s · an independent expansion strategy

Zippy’s

I revisited an introductory marketing project two years later, rebuilding the strategy around what had changed at Zippy’s.

Zippy’s plate lunch

It started as a marketing class project

A short visual overview of my independent project

2024 course project · independently revisited in 2026

Explore the full project
Currently building13 seasons backfilled · Elo validated against FiveThirtyEight
A closer lookClose preview

What is a win worth in ticket sales?

Fan Demand Lab · an NBA attendance model

A SQL-first model of NBA home attendance, built on AWS: S3 and Lambda collectors, DuckDB feature engineering, and a self-built Elo rating system validated against FiveThirtyEight's historical ratings.

13 seasons backfilled · Elo validated against FiveThirtyEight

Explore the full project
The Oops, All In logo: a poker chip tipping over with a surprised face, next to the words oops, all in. AI-assisted design. A closer lookClose preview

How confident are you in your next round?

Oops, All In · a multiplayer party game

A browser party game built with Codex and Claude Code, with me directing the design and playtesting.

How good is your internal clock?

Stop at exactly 5 seconds, without seeing the timer

Ready when you areSolo practice · aim for the smallest difference
Play Oops, All In ↗

9 minigames · up to 12 players · room-code multiplayer

Explore the full project
Tableau dashboard comparing Airbnb and hotel pricing across the Hawaiian islands, with island density maps and average price by island and neighborhood. A closer lookClose preview

Where do visitors find value in Hawai‘i?

29,011 Airbnb listings against a hotel benchmark

A Tableau dashboard comparing Hawai‘i Airbnb and hotel value, built with a course partner.

29,011 listings analyzed · 30 neighborhoods · 14-worksheet workbook

Explore the full project
Case study · in progress25,000+ listings analyzed
A closer lookClose preview

What sets prices in Tokyo’s Airbnb market?

25,000+ listings across Tokyo submarkets

An analysis of Tokyo Airbnb pricing and neighborhood differences using Python and Tableau.

25,000+ listings analyzed

Explore the full project
How I use AI

AI growth has been moving fast. It feels like there is always something new coming out. I never thought I would be able to build a web app, a video game, or even a website, but AI gave me the confidence to try. I take inspiration from what I have seen and experienced, especially problems I have run into or things that could work better. I learn by building different versions and getting a feel for what works. One of my strengths is noticing when an output feels off, then figuring out what needs to change. Curiosity keeps me learning, and I want to turn that into useful things with other people.

  1. Define the problem

    Start with the person and the decision.

  2. Explore the options

    Bring context. Compare possible approaches.

  3. Build a first version

    Make something concrete to evaluate.

  4. Check and improve

    Test the result against the original goal.

  5. Make the next pass easier

    Save what works. Automate the repeat work.

See it in a project →

AI use statement

I use AI tools, including ChatGPT, Codex, Claude, and Gemini, to explore ideas, research, write and debug code, develop designs, and revise writing. This website and several projects in this portfolio were developed with AI assistance, including parts of their code, copy, and visuals.

I set the direction, bring the personal context, choose what to keep, and work through revisions. I review the output, test how things work, and check factual claims against sources before sharing. That review can still miss things, and I take responsibility for correcting errors in the work I publish.

The level of AI involvement varies by project. My standard is to explain its role where it matters, label generated visuals, credit outside sources, and distinguish my own experience from AI-assisted exploration. Using AI does not mean I built every part from scratch or that its output is automatically correct.

This statement will evolve as the tools and my approach change. My perspective keeps evolving as the tools change and I learn from using them.

Have a role, project, or idea in mind?

Open to marketing, analytics, sports, global business, and program management roles, plus freelance projects and collaborations.