10 years ago.

If i'm not wrong, this sensor outputs uT (micro Tesla) raw values for each ot the 3 axis, the unit value of the output is 0.1uT There is nothing special you can get out of a mag sensor alone, just a not reliable north position.

#define M_PI 3.14159265358979323846f

//convert to real world values
uT_X = (float)Raw_X * 0.1;
uT_Y = (float)Raw_Y * 0.1;
uT_Z = (float)Raw_Z * 0.1;

//get heading in radians
float heading = atan2(uT_Y, uT_X);
if(heading < 0) heading += 2*M_PI;
if(heading > 2*M_PI) heading -= 2*M_PI;

//radians to degree
printf("Heading= %4.2f ", heading * 180/ M_PI);

Then, you'll see lots of noise and a not reliable heading direction, that's because: - you have magnets near the sensor, that's called hard iron distorsion (speakers, cell phones...) - you have ferromagnetic metals around, that's calles soft-iron distorsion (pc monitor, pc case....) - your sensor is not in perfect plane position, you need an accelerometer to compensate that....then you'll see that quick movement of the accelerometer will make wrong heading as "spikes", so you'll need a gyro to filter out the "shakings".

I cannot suggest a single place to get some reading, i can suggest to google for soft-iron, hard-iron, tilt compensated compass, AHRS, IMU. Anyway a single magnetometer is not enought to have a good heading.