statistics

Overview

The statistics node computes rolling statistical measures over a sliding window of incoming numeric values. Calculates mean, median, mode, standard deviation, min/max, sum, count, and percentiles. Essential for real-time sensor monitoring, anomaly detection, trend analysis, and data quality assessment on edge devices.

Properties

Statistical Operations

mean:   arithmetic average
median: middle value (sorted)
mode:   most frequent value

Window: [22, 23, 23, 24, 25]
mean:   23.4
median: 23
mode:   23

stddev: standard deviation
min:    minimum value
max:    maximum value

Window: [22, 23, 23, 24, 25]
stddev: 1.02
min:    22
max:    25

sum:        total of all values
count:      number in window
percentile: value at Nth %

Window: [22, 23, 23, 24, 25]
sum:    117
count:  5
p95:    25

Example: Rolling Average of Temperature

Smooth noisy temperature readings using a rolling mean over the last 20 values.

// Statistics node configuration
{
  "windowSize": 20,
  "operations": ["mean", "min", "max"],
  "property": "payload",
  "septopics": true
}

// Flow: DHT22 sensor → statistics → dashboard gauge

// Incoming readings (noisy):
// 24.1, 24.8, 23.9, 24.5, 25.1, 24.3, 24.7, ...

// Output after 20 readings:
{
  "payload": {
    "mean": 24.42,
    "min": 23.8,
    "max": 25.1
  },
  "topic": "sensor/kitchen"
}

// The rolling mean smooths noise:
// Raw:     24.1, 24.8, 23.9, 25.1, 24.3, ...
// Average: 24.3, 24.4, 24.3, 24.4, 24.4, ...
// Much smoother for dashboard display

Example: Anomaly Detection with Standard Deviation

Flag sensor readings that deviate more than 2 standard deviations from the mean.

// Statistics node configuration
{
  "windowSize": 50,
  "operations": ["mean", "stddev"],
  "property": "payload.value"
}

// Flow: sensor → statistics → function (detect) → switch → alert

// Function node: Check for anomalies
var stats = msg.payload;
var current = msg.originalValue;
var zScore = Math.abs(current - stats.mean) / stats.stddev;

msg.payload = {
    value: current,
    mean: stats.mean,
    stddev: stats.stddev,
    zScore: zScore,
    isAnomaly: zScore > 2.0   // Beyond 2 sigma
};
return msg;

// Normal reading: value=24.5, mean=24.3, stddev=0.8
//   zScore = |24.5 - 24.3| / 0.8 = 0.25 → OK
//
// Anomaly: value=28.0, mean=24.3, stddev=0.8
//   zScore = |28.0 - 24.3| / 0.8 = 4.63 → ALERT

Common Use Cases