TwiddleLab / Artificial Intelligence / Learning From Data

Data Is a Table

Lesson 1 of 11 · 5 min · Free, no account needed

What you will learn

• A machine learns from a dataset: a table of examples.
• Each row is one example. Each column is a feature, one thing you measured.
• The label column holds the answer the machine should learn to give.
• A wrong value in the table is a wrong lesson. Check your data.

Start the lesson

How the lesson goes

1. A new critter turned up at the research station. What should you write down about it?

   A machine can't look at a critter. It only gets what you write down. Tap every detail you'd record, then Check.

2. Measure two critters.

   Drag the violet jaw of the ruler to the tip of each ear and tail. Every reading drops into the table.

3. Which columns help tell Puffs from Zips?

   Drag each column head into a tray, or tap it and then a tray. Then Check.

4. The table has names for its parts.

   Tap to reveal them.

   Each row is one example: one critter.

   Each column you measured is a feature: one number or fact about every example.

   The answer column is the label. Here it's Kind.

   All the rows together are the dataset. It's all the machine will ever know.

5. Typo hunt: three values here can't be true.

   Someone copied the notebook in a hurry. Tap each impossible value. A wrong tap costs a life.

6. A different app, a different table. Which column is the label?

   Tap a column head, then Check.

7. Write down the new critter.

   Drag each measurement into its cell in the empty row. Then Check.

Next lesson: Ask Your Neighbors: kNN · Part of the course: Learning From Data

Lessons are interactive and need JavaScript. Turn it on to play them.