Scenarios for what-if analysis

In this section, we explore how to use scenarios for what-if analysis in Taipy. What-if analysis is all about exploring different scenarios to see how variable changes affect outcomes. It's essentially asking "What if?" questions and examining the results.

What is a Scenario?

A Scenario in Taipy represents an instance of a business problem. Among other features, a scenario owns some data nodes. Two different scenarios can own two alternative data nodes (from the same DataNodeConfig). This is particularly handy when you want to represent different situations, assumptions or conditions, and compare the data between them.

In addition to holding data nodes, scenarios can be attached to a Cycle to manage recurring business cases. This is ideal for time-based analyses like monthly forecasts, weekly results, daily reports, etc.

For more details on scenarios, please refer to the scenario and data management page.

Why use Scenarios?

The main advantages of using scenarios for what-if analysis are:

• Informed Decision-Making: Understand the implications of different situations. Instantiating multiple scenarios with alternative data nodes from the same DataNodeConfig allows you to manage different versions of the same dataset or parameter set. It is essential for use cases where experiments need to be conducted with different assumptions, or when data needs to be preserved for auditing and comparison purposes.
  For more details, see the multiple scenarios page.

• Temporal Analysis: Use cycles and data node scope for recurring business cases, ideal for time-based analyses like weekly forecasts, monthly planning, etc.
  For more details, see the recurrent scenarios page.

• Easy to use: Easily manage and compare multiple scenarios for various business situations.
  For more details, see the scenario comparison page.

How to use Scenarios?

What-if analysis in Taipy involves configuring data nodes, setting up scenario configurations, and instantiating new scenarios at runtime. Each scenario holding an alternative data node, represents a different situation, assumption or condition, and as an end-user you can compare the different scenarios to understand the impact of various decisions, or events on your data.

Here are the main principle:

1. Configuring data nodes: The first step consists in configuring the data nodes that will be instantiated along with the scenarios. Data nodes are the building blocks of scenarios. They represent any data such as input datasets, parameters, intermediate data, kpis, results or output dataset. They are instantiated from DataNodeConfig objects.
   For more details on data nodes, see the data integration page.

2. Configuring Scenarios: A ScenarioConfig is a configuration object used to instantiate scenarios. It encapsulates scenario-specific settings such as data node configurations, task configurations, scenario comparators, or frequency for cycle management.
   For more details on how to configure scenarios, see the scenario configuration page.

3. Instantiating multiple scenarios with alternative data nodes: At runtime, an end-user can create as many new scenarios as he/she wants from the same ScenarioConfig. Each one contains its own set of data nodes.
   For more details and examples, see the multiple scenarios and recurrent scenarios pages.

4. Comparing scenarios: At runtime, you can compare multiple scenarios by comparing the data nodes from the same data node configuration.
   For more details and examples, see the scenario comparison page.