How to design a lowpass filter?

I would like to filter the following data:

Obviously the data shows some oscillations with a quite constant frequency. How can I design an appropriate (low pass?) filter to filter out this noise?

(Matlab 2020b)

Answer 1

[Scott MacKenzie]

You can use smoothdata as a quick-and-dirty low pass filter:

% sine wave with noise example
y1 = sin(linspace(0,20*pi,200));
y1 = y1 + rand(1,200); 
y2 = smoothdata(y1, 'movmean', 5); % through lp filter

% square wave example
y3 = repmat([1 0], 10, 10); 
y3 = y3(:);
y4 = smoothdata(y3, 'movmean', 5); % through lp filter

tiledlayout(2,2);

nexttile(1);
plot(y1);
ax = gca;
ax.YLim = [-2 2];

nexttile(3);
plot(y2);
ax = gca;
ax.YLim = [-2 2];

nexttile(2);
plot(y3);
ax = gca;
ax.YLim = [-2 2];

nexttile(4);
plot(y4);
ax = gca;
ax.YLim = [-2 2];

You can increase the 3rd argument in smoothdata to get more aggressive filtering.

[Marcel345614]

Thanks for your answer!

Does someone know an method to only filter the oscillating frequency. A moving average filter is good, but is not always what I should use.