AP CSP Unit 2 Cheat Sheet: Data

The basics

What it covers: How data is represented as bits, how it is compressed, and how programs extract information from data.

Exam weight: Big Idea 2 is about 17–22% of the AP CSP multiple-choice exam.

The big question: How is all data represented with bits, and how do we compress it and turn it into useful information?

Computational thinking practices: Computational Solution Design, Algorithms & Program Development, Abstraction, Code Analysis, Computing Innovations, and Responsible Computing.

Key topics at a glance

Bits & Bytes

A bit is a 0 or 1 — the smallest unit of data. A byte is 8 bits. All data (numbers, text, images, sound) is ultimately stored as bits.

Binary Numbers

Convert binary → decimal by adding the place values (powers of 2) where there is a 1. n bits represent 2ⁿ values. Too big for the bits → overflow.

Analog vs. Digital

Analog data is continuous; digital data is discrete. Sampling converts analog to digital — finer samples = more accurate.

Data Abstraction

Bits are an abstraction: the same sequence can represent a number, letter, color, or sound depending on how a program interprets it.

Data Compression

Lossless: no data lost, fully recoverable (ZIP, PNG). Lossy: some data permanently discarded, smaller size (JPEG, MP3).

Compression Tradeoff

Lossless keeps all quality but shrinks less; lossy shrinks more but loses fidelity. Use lossless when every detail matters.

Information & Metadata

Data = raw facts; information = insight from analyzing data. Metadata = data about data (date, size, location).

Extracting & Using Data

Find patterns by filtering, sorting, and aggregating. Correlation ≠ causation. Clean data first. Programs process large datasets at scale.

The key terms you must know

Bit — a single binary digit, 0 or 1; the smallest unit of data.
Byte — a group of 8 bits.
Binary — the base-2 number system computers use.
2ⁿ rule — n bits can represent 2ⁿ different values.
Overflow — a value too large for the available bits.
Analog vs. digital — continuous data vs. discrete, bit-based data.
Sampling — measuring an analog signal at intervals to digitize it.
Abstraction (of bits) — the same bits can mean different things depending on interpretation.
Data compression — reducing the number of bits needed to store/transmit data.
Lossless compression — reduces size with no data lost (recoverable).
Lossy compression — reduces size by discarding some data (not fully recoverable).
Data vs. information — raw facts vs. insight gained from analysis.
Metadata — data about data.
Correlation vs. causation — association does not prove cause.
Cleaning data — fixing/removing errors before analysis.

Key themes to remember

Everything is bits. Numbers, text, images, and sound are all stored as 0s and 1s.
Bits are an abstraction. Meaning depends on interpretation, not the bits themselves.
Compression trades size for fidelity. Choose lossless or lossy based on whether loss is acceptable.
Data is not information until it is analyzed. Programs turn raw data into insight.
Be careful with conclusions. Correlation is not causation, and dirty data leads to wrong results.

Common exam traps

n bits = 2ⁿ values, not n or 2n. Each added bit doubles the number of possible values.
Lossy loss is permanent. You cannot recover the original data from a lossy-compressed file.
Use lossless for text and program files; lossy is for media where some quality loss is acceptable.
Correlation does not imply causation. A pattern in data is not proof that one thing causes another.
Metadata is data about data, not the main content itself (e.g., a photo's date, not the photo).
Overflow is a data-representation limit, not a compression method.
Cleaning data matters. Analyzing messy or biased data produces misleading information.