Air-Math Software EMA Implementation

[Madacool01]

While developing my Air-Math Project ( #3 ), I noticed a flaw in my tracking. Due to the natural shake of the human hand, the lines drawn were very jittery and inaccurate. After some research, I found a simple yet effective solution to the problem, known as Exponential Moving Average (EMA) Smoothing, which is mathematically defined as:
$S_{t} = \alpha X_{1}+(1-\alpha)\cdot S_{t-1}.$
Where:
$X_{1}=$ Current X/Y coordinate
$S_{t}=$ Smoothed value
$S_{t-1}=$ Previous smoothed value
$\alpha=$ Smoothing Factor $\to$ 0.1 = Very smooth/Unresponsive; 0.9 = Very fast, Highly responsive.
This allowed me to be able to smoothen the values out and get a responsive yet smooth movement.
Python Implementation:

self.smooth_x, self.smooth_y = 0, 0
self.smoothing_factor = 0.5 # EMA factor (0.1 very smooth/unresponsive, 0.9 fast/responsive
...
index_x = int(hand_landmarks[8].x * w)
index_y = int(hand_landmarks[8].y * h)

#EMA Smoothing
if self.smooth_x == 0 and self.smooth_y == 0:
    self.smooth_x, self.smooth_y = index_x, index_y
else:
    self.smooth_x = int(self.smoothing_factor * index_x + (1 - self.smoothing_factor) * self.smooth_x)
    self.smooth_y = int(self.smoothing_factor * index_y + (1 - self.smoothing_factor) * self.smooth_y)

[Madacool01]

Here is the result of me trying to write $\int x^{2}dx$ using the Software after applying the EMA-Smoothing (with $\alpha = 0.2$):

And here I am without the smoothing:

Pretty significant difference, isn't it?!