AP® Computer Science Principles · Big Idea 5 · Topic 5.6
Aggregating Data to Deduce Private Facts
Aggregation means combining separate pieces of data that seem harmless alone into a profile that identifies a specific person. Cross-referencing an "anonymized" data set with a second public data set can undo the anonymity entirely.
No single data point usually tells a stranger who you are. Your birth date matches thousands of people. Your postal code matches thousands more. Your gender narrows it further, but not much on its own. The problem is that these facts don't stay separate. When someone joins them together, the combination gets specific fast, because each new fact cuts the pool of possible matches down. Three weak clues can triangulate to one person even though none of the three clues is identifying by itself.
This is exactly how real re-identification works. A hospital or city agency releases a data set with names stripped out, keeping only birth date, postal code, and gender, believing that removing names made it anonymous. A researcher takes that stripped data set and cross-references it against a second data set that is public and does have names attached, like voter registration records. Matching the three shared fields across both data sets often narrows the match down to exactly one row, and that row in the voter file has a name on it. The anonymized record is anonymous no more.
The same logic scales up with location history, purchase records, and social media activity: none of these streams alone builds a full picture, but joined together they can reconstruct where someone lives, works, and what they buy. This connects directly to what metadata reveals about a file. A photo's timestamp and GPS tag seem like small technical details, but aggregated with other metadata they can expose a person's routine just as reliably as joining separate data sets can expose their identity.
Key terms
- Aggregation
- Combining multiple separate pieces of data into a single profile or conclusion.
- Re-identification
- Matching an anonymized record back to a specific named person using outside data.
- Cross-referencing
- Comparing shared fields across two different data sets to find matching records.
- Anonymized data
- A data set with direct identifiers like names removed, intended to protect individual identity.
- Metadata
- Data that describes other data, such as a file's timestamp or location, rather than its content.
Watch it explained
- How it worksHow does joining location, purchase, and social data build a profile nobody handed over?
- Worked exampleShow me one person picked out of a crowd from a birth date, a postal code, and a gender.
- Going furtherWhat changes about anonymized records once they are cross referenced with a public data set?
- Ties togetherHow does aggregating personal data connect to what metadata gives away about a file?
Check yourself
End of topic.