AI Scoring System for Artist Opportunities
Budget: $30 – $250 USD
Every Friday a scraper delivers around 350 fresh rows of residency calls, open-call exhibitions, grants, and commissions into my master spreadsheet. I need an AI layer that can read each record, weigh its value to practicing visual artists, and return a clear score with a line or two of reasoning.
The model must look beyond the obvious and judge each opportunity on: relevance to the artist’s medium, geographical fit, how close the application deadline is, the overall quality and reputation of the host organisation, the concrete benefits and provenance offered, notoriety or press potential, and any other factor an experienced artist management agency would naturally weigh.
Scope of work
• Analyse a representative sample of my existing data (I’ll share a CSV and field descriptions).
• Design a reusable prompt—compatible with ChatGPT API or equivalent—that accepts a row’s values as parameters and outputs a numeric score plus short justification.
• Deliver a weighting strategy or rubric I can adjust as priorities shift.
• Provide implementation notes so my scraping assistant can batch-process new rows every week without extra engineering.
Acceptance criteria
1. Prompt processes sample file end-to-end and returns JSON or CSV with “Score” and “Rationale” columns.
2. Weighting logic is documented clearly enough for a non-coder to tweak.
3. Accuracy spot-checked on at least 50 rows shows sensible, differentiated scores aligned with the criteria above.
If NLP libraries, Python, or Airtable scripts make things easier, feel free to incorporate them—just keep the setup lightweight and repeatable.
The model must look beyond the obvious and judge each opportunity on: relevance to the artist’s medium, geographical fit, how close the application deadline is, the overall quality and reputation of the host organisation, the concrete benefits and provenance offered, notoriety or press potential, and any other factor an experienced artist management agency would naturally weigh.
Scope of work
• Analyse a representative sample of my existing data (I’ll share a CSV and field descriptions).
• Design a reusable prompt—compatible with ChatGPT API or equivalent—that accepts a row’s values as parameters and outputs a numeric score plus short justification.
• Deliver a weighting strategy or rubric I can adjust as priorities shift.
• Provide implementation notes so my scraping assistant can batch-process new rows every week without extra engineering.
Acceptance criteria
1. Prompt processes sample file end-to-end and returns JSON or CSV with “Score” and “Rationale” columns.
2. Weighting logic is documented clearly enough for a non-coder to tweak.
3. Accuracy spot-checked on at least 50 rows shows sensible, differentiated scores aligned with the criteria above.
If NLP libraries, Python, or Airtable scripts make things easier, feel free to incorporate them—just keep the setup lightweight and repeatable.