From Data to Information to Knowledge

Raw data by itself tells you almost nothing. A spreadsheet with 10,000 rows of numbers is just data — until a program processes it and reveals a pattern. That pattern is information: facts and patterns extracted from data.

The College Board distinguishes clearly between data (raw values) and information (what you can extract from data). Data provides opportunities for identifying trends, making connections, and addressing problems — but only when properly processed and analyzed.

Often, a single data source isn't enough to draw a conclusion. Good analysis combines data from multiple sources to support a finding. A study on smartphone usage and teen anxiety might pull from phone screen time logs, clinical assessments, and survey responses — no single source tells the whole story.

Correlation vs. Causation

This is the highest-tested concept in Topic 2.3. Understand it cold.

Correlation: Two variables move together in a predictable pattern. When one increases, the other tends to increase (positive correlation) or decrease (negative correlation).

Causation: One variable directly causes changes in another.

The critical rule from the CED: A correlation found in data does not necessarily indicate that a causal relationship exists. Additional research is needed to understand the exact nature of the relationship.

Classic examples the exam loves:

• Ice cream sales and drowning rates are correlated (both peak in summer). Ice cream doesn't cause drowning — hot weather causes both.
• Shoe size and reading ability are correlated in children. Larger feet don't cause better reading — age causes both.
• A study finds students who eat breakfast score higher on tests. This shows correlation. It does not prove breakfast causes higher scores (parental income, sleep habits, and other factors may explain both).

Metadata

Metadata is data about data. Every digital file carries metadata that describes it without being the primary content.

Examples:

• A photo's primary data: the image pixels. Its metadata: date taken, GPS coordinates, camera model, file size, exposure settings.
• An email's primary data: the message text. Its metadata: sender, recipient, timestamp, subject, IP addresses of servers it passed through.
• A Word document's primary data: the text. Its metadata: author, creation date, revision history, word count.

Key AP facts about metadata:

• Metadata is used for finding, organizing, and managing information
• Metadata can increase the effective use of data by providing additional context
• Changes and deletions made to metadata do NOT change the primary data
• Metadata has significant privacy implications (covered more deeply in BI5)

Challenges of Processing Data

Real-world data is messy. The CED identifies several challenges that arise regardless of dataset size:

• Incomplete data: missing values in a dataset (some survey respondents skip questions)
• Invalid data: values that are logically impossible (age entered as -5, zip code entered as 99999)
• Non-uniform data: the same value entered differently by different people (“NY”, “New York”, “new york”, “N.Y.” all mean the same thing)
• Need to combine sources: no single dataset contains all the information needed for analysis

Data cleaning is the process of making data uniform without changing its meaning — replacing equivalent abbreviations, standardizing capitalization, removing invalid entries. It's tedious, time-consuming, and essential before any meaningful analysis can happen.

Bias in datasets: If data is collected in a biased way (from non-representative samples, with leading questions, or from a single source), the conclusions will also be biased. Critically: collecting more data does not eliminate bias if the collection method is biased. A survey of 1 million people who all use the same social media platform doesn't represent all people.

Key Vocabulary