I have a simple question, where would be the most reliable spot to install an oil sampling port for a IR 250hp Sierra compressor. I know it best before the filter, but what other criteria might I use aside from that.

When looking for a spot to locate your sample valve find a location where the oil flow is turbulant such as an elbow where the oil is changing directions this will help ensure you have a well mixed sample of whats in the system. In a straight piping run particles may settle to the bottom of the oil stream and laminar flow properties in the pipe may not give you the best results.

If you have turbulent flow "speeding" around an elbow, wouldn't you get particle size separation due to centrifugal forces? Seems to me in a past life when I designed a system for particulate sampling the rule of thumb was to place the sample point 5-10 pipe diameters from an elbow to avoid particle stratification.

From simple questions sometimes come less than simple answeres.
With that out of the way, the most reliable spot will be the spot that will consistently provide you with the data stream that holds your interest. So if your interest is in looking for wear material then as Vickie suggests select a point in the system where the system generates some selective seperation for you. Remember you are trending - looking for rates of wear particle increasses.
If you are looking for chemical stability - oxidation and or AN etc then the location doesn't matter as long as there is reasonable agitation in the system.
Define your goals(condition monitoring purpose) - most prominant failure contributor, or anticipated failure contributor. Using that as your criteria will help you to narrow down your sampling point.
regards.....
Alan

Thanks for the info, it has been very helpful.