Rotating plate thrown parallel to ground

[FGD]

[Image below]
I'm trying to figure out how much information I can get from and accelerometer and gyroscope attached to a spinning plate that is thrown through the air. It has both linear and angular acceleration/velocity. The plate is hand thrown fairly parallel to the ground. Z axis points in the direction of gravity and is the main axis of rotation. X and Y axis are close to the centre of rotation, but I am not sure exactly the coordinate of the offset of x and y.

A couple things I am confused about.

Gyroscope data in rad/s
1. I was not expecting the sinusoidal data in the gyro x, y axis since it is mostly rotating around the z axis. (What is causing the sinusoidal pattern and what does the amplitude of 10 mean?)

Accelerometer in m/s^2
1. The x an y axis sinusoidal period seems to match up to the gyro x,y period. I was expecting the sinusoidal pattern in this data. (What does the amplitude of the sinusoidal pattern of this data represent?)

2. Since the centripetal acceleration and tangential acceleration should be constant (in this case) does the dc offset of the sqrt(x^2 + y^2) equal both centripetal and tangential acceleration? If so, is there a way to separate them?

3 With z axis it also has a sinusoidal pattern which I suppose is the plate wobbling a bit as it spins. What does the amplitude of this signal represent?

So far all I can accurately get is the angular velocity, direction, and maybe a tilt based on the average of z axis acceleration (gravity), but no idea which way it is tilted.

Is there a way to extract linear acceleration/velocity and maybe orientation from this data? If so, how would you go about this?
Is there other information that I am missing that could be extracted from the data?

I attached an image of real data from this experiment.

Thank you in advance!

 

[berkeman]

What does your baseline data look like? By that I mean what do you get when you set the object down on a spinning turntable that is not translating linearly? What do you get when you set the object down on a car or motorcycle and accelerate it linearly without any spinning?

Have you considered also gathering data optically by imaging the motion of the thrown object (after adding some optical markers to the object)?

 

[FGD]

What does your baseline data look like? By that I mean what do you get when you set the object down on a spinning turntable that is not translating linearly? What do you get when you set the object down on a car or motorcycle and accelerate it linearly without any spinning?

Have you considered also gathering data optically by imaging the motion of the thrown object (after adding some optical markers to the object)?

I attached the sensor to a drill. So this is almost all rotation. You can see the centripetal acceleration that looks very similar to linear acceleration in the x, y accel axis. Z axis is close to gravity. I'm not sure why but the x,y gyro also drifts weirdly.

For linear acceleration it basically just moves up or down with acceleration, here is data from me just walking around with the plate. I did not walk just straight. But if that is needed I will get more data.