Normal Data Distribution

The normal distribution (gauss) is the most common form of distribution in the real world, shaped like a symmetrical bell around the mean. NumPy creates the data with np.random.normal(loc, scale, size), where loc is mean and scale is standard cross-section. In a normal distribution, about 68 percent of the data is in one standard crossing of the mean, 95 percent in two standard crossing. Many statistical and machine learning methods assume that data is normal.