Looking for a clever way to get the average value of a measurement

I am measuring the velocity v and the acceleration delta v / t. For the velocity, I collect samples for 100 ms before displaying the average of those samples. For the acceleration however I only store the last value for v in order to get delta v. The result is that the displayed acceleration value is spiky and unreliable.

static float v, prev_v;
static unsigned int n;
static unsigned long lastTime;

v += measure();
n++;

if((millis() - lastTime) > 100){
  Serial.print(" v = ");
  Serial.println(v / n);
  Serial.print("a = ");
  Serial.println(v / n - prev_v);
  prev_v = v / n;
  v = 0;
  n = 0;
  lastTime = millis();
}

Is there a way to also get the average value for the acceleration?

Any tips are welcome!

moving average filer

Taking the derivative increases the noise.

You could calculate the "instantaneous" acceleration from successive values of v and average that in the same way as you do v.

simple Kalman filter