Designing a Workflow

Examples

• X is conducting a survey
  • Site must be selected from pre-approved list (stored in database)
  • Snail samples must be collected at predetermined grid points (hm… where are these?)
  • GPS coordinates must be recorded for each sample (when are these checked and how do we tell if they're valid?)
  • Each sample must have a unique identifier (does the collector generate this or does the system?)
  • Snail size must be measured and recorded accurately (what are reasonable bounds for error checking?)
  • Genetic material must be extracted and properly stored (how do we know?)
  • Pollution levels must be measured at each grid point (as above…)
  • All data must be recorded digitally in the field (realistically no way to enforce scientific honesty)
• To implement survey data collection
  • Create a conduct_survey page with a multi-step form
  • Use GPS integration to automatically record coordinates
  • Implement barcode scanning for sample identification
  • Ensure offline functionality for field use
  • Implement data validation before submission
• To implement survey data upload
  • Create an upload_survey page for batch data upload
  • Check for data completeness and format consistency
  • Validate sample IDs against existing database entries
  • Ensure no duplicate entries are created
  • Use transaction to ensure all-or-nothing data insertion
  • Generate summary report after successful upload
• To implement survey data quality control
  • Add a review_survey page for supervisors
  • Allow flagging of suspicious data points
  • Implement a comment system for discussing data issues
  • Require supervisor approval before data is marked as verified
  • Log all changes and approvals for audit purposes
• X is generating reports and visualizations
  • User must have access to the data they want to report on
  • Data must be filtered according to user-selected criteria (e.g., date range, site, pollution level)
  • Reports should include summary statistics, graphs, and tables
  • User must select export format (PDF, CSV, Excel)
  • Data visualizations must be interactive and responsive
  • Only authorized users can provide feedback on reports
• To implement report and visualization creation
  • Add a create_report page with data selection options
  • Must handle user authentication and authorization checks
  • Allow users to apply filters (e.g., date range, site, pollution level)
  • Use a JavaScript library like Plotly for interactive visuals
  • Implement responsive design for all visualizations
  • Store user feedback in the database

In Detail

• Design plates for an experiment
  • Sample must not already be claimed by an existing experiment
  • Plates must parse correctly
  • Plates cannot be duplicates
  • One or more people must be associated with experiment
  • Does the uploader have to be one of those people?
• Invalidate a plate
  • Person must have privileges
  • A second person with privileges must OK the invalidation
  • An example of work in progress
• To implement experiment creation
  • Add a create_experiment page with a form
  • Must handle multiple file upload
  • And check multiple people (no duplicates)
  • No schema changes required
• To implement plate invalidation
  • Add a confirmed_by column to invalidated?
  • But what if we decide two confirmations are needed? Or three?
  • Would need a join table like performed
  • Save that complexity until the day…
  • For now just used confirmed_by
    • Display confirmations with NULL as "awaiting"