It's a pretty difficult task and I've never heard of it being done.  And because in the difficulty of duplication, I'm not sure how much precision will result.  For background, electrical resistance strain gauges work because a conductor undergoes a slight change in resistance as it is extended.  When secured to the item being tested, The wire will stretch/compress when the testing item is strained.  The change in resistance is measured using sensitive equipment, or a circuit called a Wheatstone Bridge circuit. To correct for lack of alignment, etc. strain gauges are often placed in a pattern.

To manufacture one, a length of very fine conductive wire is required.  This has to be laid out in a 'zig-zag' (like a multiple 'S' but with numerous back and forth wires. This permits a length of wire to be placed in a much shorter length. This is where the bigger problem will occur because the bends in the wire are at the ends, where there is little strain; bending the wire increases the resistance.  This wire assembly must then be transferred to the object being tested.  As a guess, the area on the tested object can have a 'glue patch' and the strain gauge can be transferred on adhesive tape.  It won't be an easy undertaking.

Why would you want to make one? Some sort of academic project? If so then contacting a manufacturer will probably get you the info you need.  I am most familiar with MicroMeasurements, and they have a great web site explaining all kinds of things regarding strain gage technology and use.  Not sure their web address but I am sure a search would find it.  

When performing an evaluation using strain gages there is no substitute for quality gages, and today's suppliers produce extremely accurate gages.  I would expect that most people could not get anywhere near the same level of quality, and any "home-made" gages would result in too much background signal noise to have much use.