Learn Sliding Window Algorithms for Streaming Analytics | Probabilistic & Streaming Data Structures

In the world of real-time data, you often need to analyze only the most recent information rather than the entire history. This is where sliding window algorithms become essential. Sliding windows enable you to focus on a fixed-size segment of the incoming data stream, which is especially important in monitoring and alerting scenarios. For example, when tracking website traffic, you might want to know the average number of users over the last five minutes, rather than since the site launched. This approach allows you to react quickly to sudden changes, detect anomalies, and avoid being overwhelmed by irrelevant historical data.


              1234567891011121314151617181920212223242526
            
from collections import deque

class SlidingWindowMovingAverage:
    def __init__(self, window_size):
        self.window_size = window_size
        self.window = deque()
        self.sum = 0.0

    def add(self, value):
        self.window.append(value)
        self.sum += value
        if len(self.window) > self.window_size:
            removed = self.window.popleft()
            self.sum -= removed

    def average(self):
        if not self.window:
            return 0.0
        return self.sum / len(self.window)

# Example usage:
window = SlidingWindowMovingAverage(window_size=3)
data_stream = [10, 20, 30, 40, 50]
for value in data_stream:
    window.add(value)
    print(f"Current window: {list(window.window)}, Moving Average: {window.average():.2f}")

There are several common types of windowing strategies used in streaming analytics, each suited to different use cases. A tumbling window divides the data stream into non-overlapping, contiguous chunks of fixed size — imagine counting logins every five minutes, resetting the count at each interval. A hopping window allows for overlapping intervals, where each window "hops" forward by a set step; for example, you might compute a summary every minute for the past five minutes, producing overlapping results. Finally, a sliding window continuously moves forward with each new data point, always containing the most recent data within its size; this is ideal for real-time averages or alerting on sudden spikes. Choosing the right window type depends on your analytics goals and the nature of your data stream.

Study More

Explore how sliding windows are used in log analytics platforms and stream processing engines like Apache Flink, Spark Streaming, and Kafka Streams to power dashboards, anomaly detection, and real-time metrics.

Everything was clear?

Thanks for your feedback!

Section 3. Chapter 4

Ask AI

Ask anything or try one of the suggested questions to begin our chat

Swipe to show menu


              1234567891011121314151617181920212223242526
            
from collections import deque

class SlidingWindowMovingAverage:
    def __init__(self, window_size):
        self.window_size = window_size
        self.window = deque()
        self.sum = 0.0

    def add(self, value):
        self.window.append(value)
        self.sum += value
        if len(self.window) > self.window_size:
            removed = self.window.popleft()
            self.sum -= removed

    def average(self):
        if not self.window:
            return 0.0
        return self.sum / len(self.window)

# Example usage:
window = SlidingWindowMovingAverage(window_size=3)
data_stream = [10, 20, 30, 40, 50]
for value in data_stream:
    window.add(value)
    print(f"Current window: {list(window.window)}, Moving Average: {window.average():.2f}")

Study More

Everything was clear?

Thanks for your feedback!

Section 3. Chapter 4