NumPy dtype: types, sizes and ranges Quick answer: A NumPy dtype describes how each element of an ndarray is stored and interpreted. Specify dtype=np.int32 for a fixed-width signed integer or dtype=np.float64 for double-precision floating point. Choose the type based on range, precision, memory and compatibility—not an assumed speedup. For example, np.array([1, 2, 3], dtype=np.int32).dtype is int32. NumPy can infer a dtype when none is supplied, but the inferred type can… Read More »NumPy dtypes Explained: Integer, Float, Boolean and String Types