Count Engagements - Aggregate User Interaction Events Within Time Windows

Question

## Round 1 - Coding

## Problem

You receive a stream of user engagement events, each with a user ID, event type, and timestamp. Given a list of queries, each asking for the total engagement count for a user within a time range, answer all queries efficiently.

```python
def count_engagements(
    events: list[tuple[str, str, int]],  # (user_id, event_type, timestamp)
    queries: list[tuple[str, int, int]]  # (user_id, start_time, end_time)
) -> list[int]:
    # For each query, return count of events for that user in [start, end]
    ...
```

```
Example:
events = [
    ("u1", "click", 10),
    ("u1", "view",  15),
    ("u2", "click", 12),
    ("u1", "like",  20),
]
queries = [
    ("u1", 10, 16),  # -> 2 (click + view)
    ("u1", 10, 25),  # -> 3 (all)
    ("u2",  0, 20),  # -> 1
]
count_engagements(events, queries) -> [2, 3, 1]
```

## Follow-ups

1. What data structure gives O(log n) per query after preprocessing?
2. How would you handle 10M events per day - batch vs. streaming approaches?
3. What if queries also filter by event type?
4. How would you handle out-of-order event arrivals in a real-time system?

Count Engagements - Aggregate User Interaction Events Within Time Windows

Question Details

Full Details

Topics

More from Pinterest