Baking is basically predictive analytics
If you’ve ever baked holiday cookies without a scale, you know the chaos of unmeasured data. A “cup of flour” is vague, the baking powder might be ancient, and the results are anyone’s guess. “Baking is basically predictive analytics,” said my colleague, Aline Bessa.
The recipe is your model, she told me, predicting a cake from the raw ingredients. Change one ingredient and the result shifts (too little fat, the dough collapses). One altered feature, and your model behaves differently.
But where baking really parallels data science is in the orchestration. Data science workflows break tasks into clear modular steps, just as a kitchen assigns mixing, blending, testing, and evaluating to different cooks. Agents take on parallel tasks (setting the table, buying napkins, inviting friends).
Multiple processes, orchestrated by you, the data expert = an accurately predicted outcome aka good cookies and Happy Holidays.
Q&A: How AI is reshaping skills, teams & strategies in 2026
Digital Journal interviewed Iris Adae, VP of Data & Analytics at KNIME to find out. Read about her thoughts in this new article.
Will AI go pop? Don’t rely on vibes
Aline, senior data scientist, passionate chef, and inspiration for today’s email recommends the More or Less podcast. In this episode: Can data tell if a stock market crash is coming?