Rolling average of active sensor data

Python Help
1

I am Making a scale for a project that measues every second, i want to take a rolling average of say 5 data samples and output the average instead of the active data.

import RPi.GPIO as GPIO 
import numpy
GPIO.setwarnings(False) 
GPIO.setmode(GPIO.BCM)
GPIO.setup(23, GPIO.OUT)

import time
import sys

EMULATE_HX711=False

referenceUnit = 1 

if not EMULATE_HX711:
    import RPi.GPIO as GPIO
    from hx711 import HX711
else:
    from emulated_hx711 import HX711

def cleanAndExit():
    print("Cleaning...")

    if not EMULATE_HX711:
        GPIO.cleanup()
    print("Bye!")
    sys.exit()

hx = HX711(5, 6)

hx.set_reading_format("MSB", "MSB")

hx.set_reference_unit(-212.2456) #for calibration hash out devide number by wei>
hx.reset()

hx.tare()

print("Add weight now.")
while True:
    try:
        val = hx.get_weight(5)
        print(val)
        time.sleep(0.1)

        hx.power_down()
        hx.power_up()
        time.sleep(0.1)

    except (KeyboardInterrupt, SystemExit):
        cleanAndExit()

any help would be greatly appreciated.

2

Not too sure if this is of help; it’s a simple illustration of how to output a moving average:

numbers = [number for number in range(1, 100, 3)]  # generate some data points

step = 5

for index, _ in enumerate(numbers):
    data = numbers[index:index + step]  # get step number of data points
    if len(data) == step:  # only calculate if there are step number of data points
        print(f"Average: {sum(data) / step}")  # the moving average
        print()
    else:
        break

You’ll need to apply this to your method of data point generation, but I think that in principle it should work. numbers here is a list (just a POC object), but a tuple would work just as well.

3

tuple would not work with active data and neither will the example, i don’t currently have the data being dumped to a log file and it only exists in the active terminal, is there any way average the data this way?

4

Look at collections.deque. For every sample, .popleft() then append(). Initialize it with N elements (append only) then enter your popleft()/append() loop, averaging the contents after each append.

5

Yeah, my first thought was deque, too. Pretty trivial when you use the right data structure:

from collections import deque
from statistics  import mean

values = deque([], 5)

for i in range(16):
    values.append(i)
    print(len(values), mean(values))