How to use filter correctly for vector data?

Hello, I would like to ask to this: I have data from txt file and I want to plot them, do the first and second derivatives. Because it's vector data so I used filter, but it isn't solve my problem. For example: One part of data(original plot) has a quadratic function characteristic and after the first and second derivatives it isnt true (1. linear and 2. constant), there is again, the non-smooth curve.

It's possible that I use filter wrong or on the wrong place?

dt = 0.000039;                                      
windowSize = 100;                                     
pocatecni_t = 22.672;                                
koncove_t = 22.736;                                 

 fileID = fopen('blabla.txt','r');
 s = textscan(fileID,'%f %f','HeaderLines','2');
             fclose(fileID);                         %scan

 y1 = s{1}*10                                         %1. laser
 y2 = s{2}*10                                         %2. laser
 t = transpose(0:dt:dt*(length(s{1})-1));             

 rozsah = (t > pocatecni_t   &   t < koncove_t);      

 %filter
 velikost_okna = 10
 b = (1/velikost_okna)*ones(1,velikost_okna);
 a = 1;

 yy1 = filter(b,a,y1);
 yy2 = filter(b,a,y2);

 %derivatives
 dy1 = diff(yy1);
 dy2 = diff(yy2);
 t1 = t(2:end)        %x
 d11 = dy1/(1e3)      %y1
 d12 = dy2/(1e3)      %y2 

 ddy1 = diff(dy1);
 ddy2 = diff(dy2);
 t2 = t(3:end)        %x
 d21 = ddy1/(1e3)     %y1
 d22 = ddy2/(1e3)     %y2

 %PLOTS
 graf = figure;

 %graf_posuvů_po_celé_délce
 subplot(4,1,1);
 plot(t,y1);
 title('Hodnoty rychlosti a zrychleni z namerenych posuvu lasery')
 xlabel('cas [t]')
 ylabel('posuv [mm]')

 subplot(4,1,2);
 plot(t(rozsah),yy1(rozsah));
 xlabel('cas [t]')
 ylabel('posuv [mm]')

 subplot(4,1,3);
 plot(t1(rozsah), d11(rozsah),'r');
 xlabel('cas [t]')
 ylabel('rychlost [m/s]')

 subplot(4,1,4);
 plot(t2(rozsah), d21(rozsah),'g');
 xlabel('cas [t]')
 ylabel('zrychleni [m/s^2]')