What is anomaly detection in property transaction data?

Anomaly detection is the use of statistics and AI to find unusual patterns in property transaction data, such as prices far from the norm, repeated sales between the same parties, or odd timing, which may signal errors, fraud, or money laundering, so reviewers can look at them.

What It Is

• The system learns what normal transactions look like, in terms of price per square foot, the parties, timing, and locations, and flags those that differ strongly, such as a flat sold at a tiny fraction of the market price, or a property that changes hands several times quickly, and sends them to analysts or officials for checking.

How It Works

• Anomalies are not proof of wrongdoing, since there can be genuine reasons, such as a family transfer or a distressed sale, so flagged cases need human review, and the quality of data is critical, as registration records can have errors or missing fields, and models can create false alarms or miss clever schemes.

What to Check

• Organisations that use such tools should clean the data, set thresholds, review flags with people, record the outcomes, retrain the models, protect personal data, and follow legal rules on reporting suspicious transactions, while buyers should note that unusually low or high prices in a deal may attract questions, and be prepared to explain.
• Document the reasons for unusual prices.

Flags need judgement, so clean the data, set sensible thresholds, review every flag with a person, and document the reasons for unusual prices.