How to Count Events by Minute with a Tumbling Window in Python

Group timestamps into fixed 60-second tumbling windows and count events per bucket using a dict.

Medium Python 3.9+ Aug 9, 2026 Data pipelines & processing 13 views 0 copies

datetime grouping time-window aggregation dict

Python code

27 lines

Python 3.9+

from collections import defaultdict
from datetime import datetime, timedelta


def tumbling_window_count(events, window_seconds=60):
    buckets = defaultdict(int)
    for event in events:
        ts = datetime.fromisoformat(event["timestamp"])
        bucket_start = ts - timedelta(seconds=ts.second % window_seconds,
                                        microseconds=ts.microsecond)
        bucket_key = bucket_start.replace(second=0, microsecond=0)
        if window_seconds > 60:
            bucket_key = bucket_start.replace(minute=bucket_start.minute - bucket_start.minute % (window_seconds // 60),
                                              second=0, microsecond=0)
        buckets[bucket_key.strftime("%Y-%m-%d %H:%M:%S")] += 1
    return dict(sorted(buckets.items()))


if __name__ == "__main__":
    events = [
        {"timestamp": "2025-01-01T10:00:00", "event": "click"},
        {"timestamp": "2025-01-01T10:00:30", "event": "click"},
        {"timestamp": "2025-01-01T10:01:45", "event": "view"},
        {"timestamp": "2025-01-01T10:02:10", "event": "click"},
        {"timestamp": "2025-01-01T10:02:55", "event": "view"},
    ]
    print(tumbling_window_count(events))

Output

stdout

{'2025-01-01 10:00:00': 2, '2025-01-01 10:01:00': 1, '2025-01-01 10:02:00': 2}

How it works

The function uses datetime.fromisoformat to parse each event's timestamp into a datetime object. It then calculates the bucket start by subtracting the seconds and microseconds past the window boundary. For the default 60-second window, replace(second=0, microsecond=0) snaps to the minute. The key is formatted as a string for easy readability and sorted for deterministic output. The defaultdict(int) automatically initializes counts to zero, simplifying the increment logic.

Common mistakes

Forgetting to reset microseconds when computing bucket start, causing off-by-one boundaries
Using `window_seconds` values that don't divide evenly into 60 for minute-based grouping
Not sorting the final dict, leading to non-deterministic output order

Variations

Use pandas `resample('T').count()` for large datasets with a DataFrame
Implement a pure numeric bucket key (epoch seconds // window) for faster comparisons

Real-world use cases

Aggregating web analytics events into per-minute traffic counts for dashboards.
Batching sensor readings into fixed time windows for downstream anomaly detection.
Grouping payment transactions by minute to monitor fraud spikes in real time.

How to Count Events by Minute with a Tumbling Window in Python

Python code

Output

How it works

Common mistakes

Variations

Real-world use cases

More from Data pipelines & processing

Tutorials

Quizzes

Python code

Output

How it works

Common mistakes

Variations

Real-world use cases

More from Data pipelines & processing

Keep learning

Tutorials

Quizzes