To create the nuclear bomb during the Second World War, physicists needed to perform pretty complicated calculations. Stanislaw Ulam got the idea to treat this challenge as a game of chance. Later, the method he came up with was given the code name Monte Carlo. Games of chance usually have very simple rules, but playing in an optimal way can be difficult. According to quantum mechanics, subatomic particles are also unpredictable. If we simulate many experiments with subatomic particles, we still can get an idea of how they are likely to behave. The Monte Carlo method is not deterministic, but it approaches the correct result for a complex computation for a sufficiently large number of simulations.
The statsmodels.distributions.empirical_distribution.ECDF class defines the cumulative distribution function of a data array. We can use its output to simulate a complex process. This simulation is not perfect, because we lose information in the process.