Python based algorithms can be broken into smaller tasks, which can then be reused in future projects. Each workflow can be run in multiple versions with each version having perhaps different parameters. This will allow easy comparison of different algorithm settings and choose the better one fit for your use case. Each workflow can be used triggered based on events, scheduled or run immediately as soon as data is available. This will be a powerful tool for user to create analytics that he needs to present to his clients or customers.