Skip to main content

Validation Workflow

The Gesund.ai Platform's Validation workflow is designed to ensure the accuracy and reliability of the validation process through a well-defined series of steps. By following this workflow, a systematic approach is maintained, facilitating the identification of data inconsistencies, verifying accuracy, and ensuring high-quality results.

Validation Workflow

Workflow Overview​

The validation workflow consists of the following key steps:

  1. Initialization
  2. Data Retrieval
  3. Data Processing and Filtering
  4. Data Storage
  5. Summary Generation
  6. Tracking Update
  7. Optional Metric Generation
  8. Completion and Reporting

1. Initialization​

  • The validation process is initiated either by:
    • A user action (e.g., submitting a validation request).
    • An automated event triggered by the platform based on predefined rules or conditions.

2. Data Retrieval​

  • Relevant data is retrieved from the database using specific criteria:
    • Batch Job ID: To track the specific job.
    • Meta Filters: To narrow down the data based on metadata requirements.
    • Sub-cohort ID: To focus on a particular subset of the data if needed.

3. Data Processing and Filtering​

  • The retrieved data undergoes processing and filtering:
    • The platform analyzes the data, applies necessary transformations, and performs any required calculations.
    • A validation collection is formed that matches the specified criteria and is ready for further analysis.

4. Data Storage​

  • The validation collection is saved back into the database:
    • This ensures that the processed and filtered data is accessible for future reference, audits, or additional analysis.
    • It provides a point-in-time record of the validation dataset.

5. Summary Generation​

  • A comprehensive summary report is generated:
    • This report consolidates key metrics and information regarding the validation process.
    • It highlights important details such as:
      • Main Metrics: Key performance indicators.
      • Statistical Measures: Information like mean, variance, etc.
      • Relevant Insights: Any noteworthy observations or anomalies found.

6. Tracking Update​

  • The platform updates the batch job tracking:
    • Validation status, progress, metrics, and associated metadata are all updated to reflect the current state of the validation process.
    • This ensures that the user or automated systems can monitor the validation’s progress in real-time.

7. Optional Metric Generation​

  • If additional insights are needed, the platform can generate additional metrics:
    • For example, BlindSpot may be used to identify any blind spots or anomalies within the validation data.
    • This step is optional and depends on user preferences or automated rules.

8. Completion and Reporting​

  • The validation process concludes, and the final report is prepared:
    • The results are shown on the platform's UI, including statistical tables, charts, and graphs.
    • The report provides:
      • A summary of validation results.
      • Insights derived from the analysis.
      • Any discovered issues or recommendations for further investigation.
      • Actionable Information: Relevant details for follow-up or remedial actions.

Additional Notes​

  • The validation workflow is designed to be robust, handling datasets of varying sizes and complexities.
  • Real-time tracking and status updates provide users with transparency throughout the validation process.
  • The report output is customizable, allowing users to choose specific metrics and insights they wish to highlight.