Beginner's Guide to Data

Главные выводы

• Data is any record of what happened, kept the same way every time, and it beats memory because anyone can count it and check it.
• Begin every piece of data work with one specific question tied to a decision, not with a spreadsheet full of numbers to explore.
• Keep tables tidy: one row per thing, one column per detail, one value per cell, and one fixed format for dates and names.
• Look at the median and the extreme values alongside the average, because a few unusual cases can drag an average far from typical.
• Before claiming a cause, ask what else changed, who is missing from the records and what total each count came from.

Внутри этого бита

1. What data actually is

   Data is any record of something that happened, written down in a consistent way so it can be counted, compared or checked later. Numbers are the obvious kind, but data also includes dates, product names, yes-or-no answers, ratings and short comments.

2. The words that unlock any spreadsheet

   A dataset is a collection of records about one kind of thing, usually laid out as a table. Each row is one record, sometimes called an observation: one sale, one student, one delivery.

3. First move: start with a question

   The common beginner instinct is to open a spreadsheet and hunt for something interesting. It rarely works, because a table can be sliced a hundred ways and most lead nowhere.

4. Second move: build a tidy table

   Once you know your question, gather the records into a tidy table. The rules are short.

5. Third move: summarise, then picture it

   With a tidy table you can start answering. Most beginner questions need only four operations: count how many, sort from largest to smallest, filter to the rows that matter, and summarise with an average or a median.

6. Classic beginner mistakes to skip

   Four mistakes catch almost everyone early on. The first is reading cause into a pattern.