Skip to main content

Scenario Node: AI-Powered Predictive Analysis

The ScenarioNode is an advanced predictive analysis tool that leverages an LSTM (Long Short-Term Memory) model to forecast future data trends. Its primary function is to provide insights into how different market conditions or specific events could impact your data, allowing for robust "what-if" analysis.

Core Functionalities​

  • LSTM Predictive Model
    The node automatically trains a neural network on your historical data to generate a series of future predictions. This model is adept at recognizing and extrapolating patterns in time-series data.
  • Predefined Scenarios
    A set of built-in scenarios is available to instantly simulate common market dynamics.
  • Custom Scenario
    This feature allows you to define your own events or "shocks" for unique modeling.
  • Interactive Visualization
    Results are presented in an interactive chart that separates historical from forecasted data, with summaries and key metrics.

Default Scenarios​

Normal Scenario​

Represents the baseline, most likely future based on current trends. Assumes stable market conditions. Summary shows projected percentage changes.

Volatile Scenario​

Simulates a highly unstable market with unpredictable swings and sudden shocks.

Black Swan Scenario​

Models an extremely rare catastrophic event, followed by uncertain recovery.

Crisis Scenario​

Prolonged downturn unfolding over time, useful for modeling sustained slowdowns.

Recovery Scenario​

Optimistic turnaround with accelerating growth.

Stagnation Scenario​

Economic environment with stalled growth and low activity.

Growth Scenario​

Consistently favorable conditions leading to steady expansion.

Custom Scenario: Creating Your Own Shocks​

Define events by configuring:

  • Step: Index of prediction where shock is applied (0 = first point)
  • Column: Data column affected
  • Type:
    multiplier – multiplies value
    additive – adds value
    set-value – replaces with fixed number
  • Value: Numeric value applied

Use Case 1: Simulating a Sales Collapse in a Foreign Market​

Objective: Forecast the impact of a sudden drop in sales in a specific country.
Required Dataset: Column for sales in that country (e.g., sales_india).
Shock Configuration:
Step: 2
Column: sales_india
Type: multiplier
Value: 0.5

Explanation: Models an event like a trade barrier that halves sales, projecting long-term effects and recovery.

Use Case 2: Simulating an Increase in Production Due to Investment​

Objective: Model long-term effects of a capital investment.
Required Dataset: Columns like production_volume, production_cost.
Shock Configuration:

  • Step: 4 | Column: production_volume | Type: additive | Value: 1000
  • Step: 8 | Column: production_cost | Type: multiplier | Value: 0.9

Explanation: Boosts production first, then lowers costs due to efficiency gains.

Use Case 3: Simulating a Supply Chain Disruption​

Objective: Predict effects of a raw material shortage.
Required Dataset: Columns like raw_material_supply, production_volume, production_cost.
Shock Configuration:

  • Step: 3 | Column: raw_material_supply | Type: multiplier | Value: 0.6
  • Step: 3 | Column: production_cost | Type: multiplier | Value: 1.2
  • Step: 4 | Column: production_volume | Type: multiplier | Value: 0.9

Explanation: Shortage raises costs immediately, reducing production in the next period.

Use Case 4: Simulating a Market Boom​

Objective: Forecast the impact of sudden, sustained growth.
Required Dataset: Columns like sales_global, customer_acquisition, revenue.
Shock Configuration:

  • Step: 1 | Column: sales_global | Type: multiplier | Value: 1.5
  • Step: 1 | Column: customer_acquisition | Type: additive | Value: 5000

Explanation: Models a boom from a viral launch or sentiment shift, showing combined impact on revenue.