A good example of cleaning data is in cricket. The ball speed of each delivery is recorded but the technology is not perfect and some speeds come back as impossibly high or low. These are removed and only plausible results are entered and used to calculate averages speeds and other statistics.