The amplitude was large at first.
It was still swinging after 15 minutes but the amplitude was so small that it was hard to detect much change. To be fair, I realised that I didn't know in which direction it would rotate and thought I detected a very slight clockwise rotation. It turns out that this is correct for the northern hemisphere. However, I need to get it to swing for a lot longer. My first calculation will be the drag force, which = 1/2 x drag coefficient x cross sectional area x density x speed^2. The bob was a vertical cylinder of height 3cm and diameter 2cm. I got the formula for drag coefficient from http://documentation.dsaocean.com/tutorials/Tutorials/PDS-ACP.html = 0.66 approx. Density of air is about 1 kg per cubic metre. Time period is about 3 seconds and it swung about 50cm there and 50cm back so speed = 0.33 m/s approx. Putting the numbers in gives an average drag force of order = 0.00002N. Then average work done every 3 seconds = 0.00002J. So in 15 minutes that is 0.006J. Bob has mass of about 50 grams. If the work done by air resistance has dissipated all of the gpe put in at the start, then mgh = 0.006J and thus h=0.01 metres, or 10cm. It probably wasn't raised quite that high but the calculation does suggest that air resistance is the biggest cause of slowing.