Would it be possible to use mobile phones to determine the velocity of a seismic wave? I imagine an experiment similar to the velocity of sound.
1) Place two people with hammers at a distance of 5 meters on a concrete floor
2) Place two phones between them
3) Measure the distance between the phones
4) Apply the first hammer to start the stopwatches
5) Apply the second hammer to stop the stopwatches
I like the idea, but I don't think that the temporal resolution would be sufficient at a distance of 5 meter. The trick for the speed of sound experiment is that the microphone records at 48kHz and we can easily achieve a resolution of 1ms (mathematically, this could be 21µs, but in practice that's not realistic due to finite rise times and dispersion). Across 5m you would need to measure a difference of about 30ms (forth and back 10m/(330m/s)), which is fine.
For your experiment, both, the speed of sound as well as the resolution of the sensor change for the worse. The accelerometer typically acquires data at a rate of 100 Hz. The fastest one I have ever found in a Samsung phone could do 500 Hz. (To be precise: The actual sensors usually can do more, but that's what the manufacturer exposes for several good reasons.) So, even in the best case, neglecting any practical limitations, we can resolve 2ms at best, 10ms for most devices. For concrete a quick search gave me a speed of sound of about 3500m/s, so we would need to resolve 3ms.
So, in the end you might just see that the phones measure a difference of one sample, which really is not much of a measurement. This might start to become interesting across 50m and still you need one of the phones with the rather extreme accelerometers (I only know of some Samsung flagships and the Nexus 5x). Not sure if you could still detect the motion.