Process Sensor Streams With Time Windows | Dataford Interview Questions

Cambridge Mobile Telematics DriveWell receives timestamped sensor readings during a trip. Detect dense bursts of readings that may indicate a significant driving event or sensor anomaly.

Given readings sorted by strictly increasing timestamp, emit an alert at the timestamp of the first reading that makes the trailing inclusive window contain at least min_readings readings. Emit only one alert per burst. A new alert may be emitted after the window count falls below min_readings and later reaches the threshold again.

Formal Specification

Implement detect_bursts(readings, window_seconds, min_readings). readings is a list of [timestamp, value] pairs, where timestamps and sensor values are integers. Return a list of timestamps at which alerts are emitted. The sensor value does not affect burst detection.

For a reading at time t, its window is [t - window_seconds, t], including both endpoints.

Constraints

0 <= len(readings) <= 10^5
Each reading has the form [timestamp, value].
Timestamps are strictly increasing non-negative integers.
0 <= timestamp <= 10^9
window_seconds >= 0
1 <= min_readings <= len(readings) when readings is non-empty

Function Signature

def detect_bursts(readings, window_seconds, min_readings):

Your question is Process Sensor Streams With Time Windows. Start with the requirements in the Question tab.

Run and submit as often as you like. When you're ready, talk me through your approach or go straight to the code.

Problem