Data Preparation

Data preparation for cycle and chain algorithms

Effective application of the cycle and chain algorithms begins with proper data preparation. Data is the foundation upon which algorithms operate, and ensuring its accuracy, consistency, and structure is crucial for achieving meaningful outcomes.

Understanding the data structure

The cycle and chain algorithms work with data formatted in a specific way: a Comma-Separated Value (CSV) file containing three columns:

1. First Column: Identifier

• A merit order for ranking purposes.
• An order of appearance to represent the sequence.
• An ID number or any other unique identifier relevant to your context.

2. Second Column: Current Position

• A person’s current role or location in a job transfer system.
• A vehicle’s current owner in property exchanges.
• The DOI (Digital Object Identifier) of a research article in citation analysis.

3. Third Column: Expected Position

• The new role or location an individual seeks in employee transfers.
• The target owner of a property in exchange systems.
• Articles that a researcher aims to cite or build upon.

Capturing the data

1. Online Forms

• Use dropdown menus or radio buttons to minimize errors.
• Ensure uniform data entry, especially in job transfer scenarios.

2. Real-Time Applications

• Capture data as events occur for real-time accuracy.
• For example, a mobile app could record property exchanges.

3. Standalone Applications or Spreadsheets

• Useful for smaller datasets and initial experiments.
• Ensure proper data integrity maintenance.

Saving data as CSV

CSV files store data in a simple text format where columns are separated by commas. Example:

ID,Current Position,Expected Position
1,Role A,Role B
2,Role B,Role C
3,Role C,Role A

Ensuring data quality

1. Uniformity in Data Entry

• Use consistent terminology and formatting.
• Avoid inconsistencies like "Dept A" vs. "Department A".

2. Error Reduction

• Use dropdowns and radio buttons to minimize errors.

3. Validation

• Regularly review datasets for missing or duplicate entries.

4. Backup and Storage

• Keep backups to prevent data loss.
• Secure CSV files containing sensitive data.

Tailoring data collection to your use case

Examples of tailored approaches:

• Employee Transfers: Capture employee IDs, current roles, and desired roles.
• Property Exchanges: Record property identifiers, current owners, and intended buyers.
• Research Publications: Compile research article DOIs and citations.

By following these best practices, you can ensure seamless integration with the cycle and chain algorithms.