Noise in data is meaningless data or unexplained variation in data which might be due to instrument errors, corruption or other issues. Noise disguises and/or distorts the underlying data which make it harder to analyse, just as noisy environments make it more difficult to hear the sound on which you wish to focus.