What Is Data?

Introduction

Data is everywhere: the steps your phone counts, the prices in a shop, the scores in a game. Learning to understand data helps you make better decisions, spot misleading claims and, if you wish, build a career. You do not need to be a maths expert to start.

The simple explanation

Data is recorded facts. Analysis is asking questions of those facts.

One fact alone, such as "it rained on Monday", is not very useful. A year of daily weather records can answer "which month is the wettest?".

Tables: rows and columns

Most data is organised in a table, like a spreadsheet:

• Each row is one record, here one person.
• Each column is one kind of information, called a variable or field.
• Each box holds a value.

Types of data

• Numerical: amounts you can calculate with, such as age, price or temperature. Some are whole counts (number of children), and some are measurements (height).
• Categorical: labels for groups, such as city, colour or job title. You can count them but not average them.
• Ordered categories: categories with a natural order, such as small, medium and large.
• Dates and times: points in time, which allow questions about trends.
• Text and other data: reviews, images, audio. They can be analysed too, but need extra steps.

Knowing the type tells you what you can do with it. You can find the average of ages, but not the average of cities.

Where data comes from

• Collected by you: surveys, experiments, your own records
• Collected by organisations: sales, website visits, sensor readings
• Open data: public data published by governments, researchers and charities, usually free to use under a licence
• Other people's reports: which you should check carefully

Good and bad data

Data can be wrong, incomplete or unrepresentative. Typical problems:

• Missing values and typing mistakes
• Duplicates
• Different formats for the same thing ("UK", "U.K.", "United Kingdom")
• Out-of-date information
• Data that covers only some kinds of people or events

Checking and cleaning data comes before analysis. Wrong data gives wrong answers, however clever the method.

Privacy and ethics

Data about people needs care.

• Collect only what you need, and be clear why.
• Protect it, and do not share it carelessly.
• Follow privacy laws and the rules of where you work or study.
• Remove or hide names and other details that identify people when you do not need them.
• Remember that data about people can be misused or unfair, so think about who might be affected.

Try it yourself

Choose something you can record for a week, such as hours of sleep, cups of tea or minutes spent walking. Make a table with a date column and at least two other columns. Mark which columns are numerical and which are categorical.

Summary

• Data is recorded facts, usually arranged in tables of rows (records) and columns (variables).
• Types include numbers, categories, dates and text, and the type decides what you can do with it.
• Check data quality before analysis, and treat data about people with privacy and care.