A thin MEMS microphone can be used to detect audio levels and even perform basic FFT functions. You can read the analog voltage corresponding to the audio on analog pin #A4. Note that this is the raw analog audio waveform! When it's silent there will be a reading of ~330 and when loud the audio will read between 0 and 800 or so. Averaging and smoothing must be done to convert this to sound-pressure-level.
The microphone is sensitive to 100 Hz - 10,000 Hz audio frequencies. See the datasheet for more details.
You can read the raw analog value from the amplified microphone with
CircuitPlayground.soundSensor() this will give you just a value from between -1023 and 1023 where the default 'quiet' reading is 0. However, we don't really recommend using this, because its slow and gives you only a single point.
Chances are you want to measure overall volume. In which case, call
CircuitPlayground.mic.SoundPressureLevel(time) to get a "Sound Pressure Level" in dB. This will range from maybe ~50 dB (quiet) to 100dB (loud!) The time argument should be number of milliseconds to listen before performing the calculations, 10 is a good place to start.
If you do need a lot of audio values, you can call
CircuitPlayground.mic.capture(int16_t *buf, uint16_t nSamples) with a pre-allocated array of int16_t's and the number of samples to capture, this is if you want to get raw values only. On AVR Circuit Classic you'll sample at 10KHz, on Express its 22KHz