Sunbird Insights Start a conversation

Clear answers from messy data, and how sure you can be about them.

Sunbird Insights is the consulting practice of Dr. Chester Ismay, a data scientist who co-wrote ModernDive and has taught data science and agentic AI for business applications to thousands of analysts for O'Reilly, DataCamp, and Data Society. In person across Arizona, remote anywhere, and on the road when it helps: trainings so far in Australia, Europe, Canada, and across the US.

Start a conversation See selected work
Chester Ismay, smiling, in a dark jacket in front of an office window.

Courses and workshops delivered for

  • O'Reilly
  • DataCamp
  • Data Society
  • Portland State University
  • Instats
  • Udacity

What I take on

Six kinds of work, each with a fixed scope and a written result that a decision maker can read without a statistics degree.

  • Sports analytics

    Player, lineup, draft, and strategy questions answered with play-by-play and tracking data, in both R and Python, with the method and its limits written down. For teams, media, and analytics firms. Past consulting work includes the Portland Trail Blazers. The studies below are the portfolio.

    A single question answered in a few weeks, or an ongoing analytics partnership through a season.

  • Training for data teams

    Live, hands-on sessions built around your own data and tools: AI literacy for analysts, statistical thinking for decision makers, moving from spreadsheets to reproducible reports in R or Python, and a working clinic on coding agents such as Claude Code.

    Half a day to two days, in person anywhere in Arizona, on site elsewhere by arrangement, or virtual, for groups of up to about 20.

  • Coaching for individuals

    One-on-one sessions for an analyst, researcher, or career changer learning R, Python, statistics, or AI tools: a personal learning plan, review of your own code and analyses, and practice on the questions you face at work.

    Remote, in blocks of sessions, with the plan written after the first one.

  • AI agents and reproducible pipelines

    Agents that read your data and send one useful message, retrieval over your own documents, and reports that rebuild from source on a schedule. Built on your stack, tested, and handed over with documentation your team can maintain.

    A proof of concept in two to four weeks, then a production build if it earns one.

  • Fractional data lead

    A senior statistician on retainer for a team that cannot hire one yet: planning, review of the team's analyses, office hours, and a steady hand on what to measure and what to ignore.

    A set number of hours each month, with a cap in the agreement.

  • Statistical consulting and review

    Experiment design, model choice, and interpretation, or a second opinion on an analysis before it goes to a board, a funder, or a journal. The report states the uncertainty plainly and shows the code that produced every number.

    A fixed-price review, or a longer study with milestones.

Across every engagement

  • Critical thinking
  • Data wrangling
  • Data visualization
  • Reproducible workflows
  • Technical communication
  • Project planning and delivery

Tools

  • R
  • Python
  • Quarto
  • Shiny
  • Jupyter
  • SQL
  • Git
  • HuggingFace

How I work

  • Plain language, no jargon overload
  • Collaborative by default
  • Timely and organized
  • Focused on your actual question

Selected work

Each study states the question, the method, and the uncertainty, and every published number traces to committed code in R and Python.

More studies, apps, packages, and talks are on chesterismay.com.

How an engagement runs

The same four steps every time, so you know what you are buying before any work starts.

  1. A 30-minute call

    You describe the question. I say whether I am the right person for it and what it would take.

  2. A one-page scope

    Deliverables, timeline, and a fixed price or a monthly cap. No hourly surprises.

  3. The work, in the open

    Weekly check-ins, code in a shared repository, and numbers that trace to a committed source.

  4. Handoff

    A written summary for decision makers, the reproducible code, and a session with your team so the work keeps running without me.

The Sunbird Insights logo: an orange bird perched on a branch of data nodes in front of a gold sun.

About Chester

I am a data scientist and educator based in Phoenix, Arizona. My work sits where statistics, computing, and plain explanation meet: I have spent the past decade teaching analysts and researchers to reason with data, and building the tools and books they learn from.

I co-wrote Statistical Inference via Data Science (ModernDive) and maintain its R and Python packages, and I am co-writing Parallel Simulation Experiments on a Personal Computer for the Chapman and Hall/CRC Data Science Series. I teach live courses for O'Reilly, DataCamp, Data Society, Instats, and Portland State University, and I host O'Reilly's Monday Morning Agent series.

Sunbird Insights, LLC is the Arizona company through which I take consulting work. The name comes from a love of being outside: hiking and al fresco dining year round whenever possible.

  • Ph.D. in Statistics, Arizona State University
  • M.S. in Statistics, Northern Arizona University
  • B.S. in Applied and Computational Mathematics, South Dakota School of Mines and Technology
  • Co-author of the infer and moderndive R packages

Start with a message

Tell me the question you are trying to answer and the deadline. I reply within one business day with whether I can help and what the first call would cover. I work with sports teams and media, nonprofits and mission-driven groups, universities and curriculum teams, researchers, and early-stage companies that need a data lead before they can hire one.

Or write directly.

hello@sunbirds.ai