Apply Ideas¶

Technique: Compare recent behavior to a historical baseline to detect meaningful change.

Drift occurs when a system gradually changes its typical behavior.

A good dataset for this module:

• includes a historical period and a recent period
• represents repeated measurements of the same system

Example Systems¶

Website Performance¶

Possible fields:

• requests
• errors
• latency_ms

Questions to explore:

• Has error rate changed over time?
• Is average latency drifting upward?

Retail Demand¶

Possible fields:

• date
• units_sold

Questions to explore:

• Has demand shifted compared to previous weeks?
• Are there gradual changes in purchasing behavior?

Environmental Sensors¶

Possible fields:

• timestamp
• air_quality_index

Questions to explore:

• Is pollution increasing over time?
• Are recent readings different from historical patterns?

System Resource Usage¶

Possible fields:

• cpu_usage
• memory_usage

Questions to explore:

• Has average system load changed?
• Does the system appear to be drifting toward higher resource usage?

Transportation Demand¶

Possible fields:

• rides
• date

Questions to explore:

• Has ridership changed compared to earlier periods?
• Are changes gradual or sudden?