Four work products supporting the Cultural Investment Strategy.
The Duncanville Arts Foundation has built the operational infrastructure to execute, measure, and report on arts programming as economic development activity. This document describes four interconnected work products: the public-facing strategy website, the Program Manager scope of work, the CII Data System that stores and analyzes all activation performance data, and a planned AI Agent layer that will augment scoring, recommendations drafting, and research reporting within the Data System.
Why this brief exists.
The Duncanville Arts Foundation has built the operational infrastructure to execute, measure, and report on arts programming as economic development activity. This document describes that infrastructure — not to demonstrate organizational capability, but to establish that the data collection systems, measurement instruments, and documentation standards required for rigorous research are in place and functional.
The work products described here are not aspirational. The website is live. The Program Manager scope is drafted and ready for contract insertion. The CII Data System is built and awaiting its first activation record. The AI Agent layer is defined and will be developed once the first activation cycle is underway. What remains is a certificate of occupancy from the City of Duncanville and an intake cohort.
"Can arts programming structured around demand validation and behavioral substitution measurement produce reproducible evidence of entertainment spending retention at the municipal level?"
The Cultural Investment Index is designed to answer this question one activation at a time. Each scored activation produces a data point. Multiple activations produce a trajectory. A full Year One dataset produces the first structured evidence base for cultural investment in Southwest Dallas County.
The Foundation Website
The public-facing website at duncanvillearts.org communicates the Cultural Investment Strategy to four audiences simultaneously: prospective program proposers, municipal stakeholders, economic development partners, and research institutions. It functions as the primary intake channel and the canonical public statement of the Foundation's methodology.
The website presents the economic framework underpinning the strategy — a derivation of Duncanville residents' $31 million annual entertainment expenditure calculated from the Bureau of Labor Statistics Consumer Expenditure Survey (2024) applied to U.S. Census Bureau American Community Survey household count data (2019–2023). The methodology is cited and reproducible.
The site communicates the five-stage pipeline (Intake, Develop, Validate, Activate, Graduate) and the Cultural Investment Index scoring methodology in full — including the formula, all five factor weights, four-tier rubrics, and the graduation decision framework sourced from CIS v2.0, Appendix D. The CII section is reproduced from the governing document verbatim in structure, making the website a living reference for the strategy.
The Propose section explains what a prospective proposer receives from the pipeline regardless of activation outcome: five development workshops, one-on-one consulting with Foundation staff, a written analysis of their validation campaign, and a pathway to the Cultural Activation Producer Certificate. This framing positions the pipeline as a development resource rather than a gatekeeping mechanism.
A proposer intake form embedded in the site routes submissions through Formspree to the Foundation's operational email. Submitted forms trigger an automated acknowledgment email and a confirmation page that explains the next steps in the pipeline. The form captures all data fields from Appendix A of the CIS.
Program Manager Scope of Work
The Program Manager Scope of Work defines the operational role responsible for executing Cultural Investment Strategy activations at Arts Junction. It was developed as a standalone document for insertion into CIS activation contracts, ensuring execution consistency across all programs regardless of which Foundation staff member manages a given activation.
The scope defines 31 numbered responsibilities across three implementation phases. The phased structure is not an approximation — it directly mirrors the pipeline stages from the CIS, ensuring that the Program Manager's work outputs align with the data collection requirements of each pipeline stage.
The document serves a dual purpose. Within the Foundation's operations, it functions as a contract exhibit that defines the role's boundaries, authorities, and performance standards. In the context of the research partnership, it functions as an execution fidelity instrument: by standardizing what the Program Manager does at each activation, it controls a primary instrumentation threat to the CII data — the risk that data collection quality varies by individual staff member.
The performance standards section defines five thresholds: attendance data within 24 hours, survey results within 72 hours, CII scoring complete within seven calendar days, proposer feedback within 14 calendar days, and all activation documentation filed within 21 calendar days. These thresholds are sourced from the data collection protocol in Appendix E and the processing strategy in Appendix D of the CIS.
- Validate pre-commitment campaign materials against CIS standards
- Monitor commitment velocity and milestone pacing
- Coordinate sponsor agreement execution
- Confirm venue configuration and technical requirements
- Distribute substitution survey instruments (Appendix C)
- Establish adjacent business baseline measurement (4 nights)
- Verify 100% pre-commitment threshold before activation proceeds
- Confirm all data collection systems are operational
- Execute venue setup per technical rider
- Administer substitution survey at ingress and egress
- Collect ZIP code data from attendees
- Monitor and document adjacent business activity
- Record total attendance against ticketing platform
- Document any deviations from planned format
- Collect point-of-sale data from participating ORS tenants
- Submit attendance data within 24 hours
- Process survey responses within 72 hours
- Complete CII Scoring Worksheet within 7 days
- Deliver scored worksheet to proposer within 14 days
- File all documentation within 21 days
- Enter activation record into CII Data System
- Flag programs requiring development support or sunset review
- Update multi-activation trajectory tracker
The CII Data System
The CII Data System is a web-based multi-user application that stores every activation record, calculates Cultural Investment Index scores, tracks program performance across multiple activations, and exports structured datasets for research and reporting. It is the operational database underlying the Cultural Investment Strategy's measurement framework.
The system has two layers. The front end is a private web application that program managers access with individual login credentials. It contains the full CII scoring form with all five factor rubrics embedded, automatic composite score calculation, a per-program activation history with trend visualization, a cross-program dashboard, and three structured CSV export formats.
The back end is a PostgreSQL database hosted on Supabase, a managed database service. The database holds two tables: programs (one row per program in the pipeline) and activations (one row per scored activation, containing all five factor scores, observed data text fields, the composite CII, and the determination). All authentication is handled by Supabase's row-level security layer, which restricts read and write access to authenticated users only.
Data is owned by the Foundation. The PostgreSQL database exports as a standard SQL dump at any time. The CSV export function produces three files formatted for direct import into Excel or any statistical software: a full activations dataset, a programs summary, and a multi-activation CII trajectory view. These exports are the primary data transfer mechanism for external research reporting cycles.
AI Agent Layer
The Foundation is developing an AI Agent layer that integrates with the CII Data System. The agent will operate inside the scoring workflow — reading observed data, assisting with factor scores, drafting recommendations, generating narrative reports, and surfacing cross-program patterns for research analysis. All agent outputs are reviewed and approved by Foundation staff before they enter the record.
The agent connects to the same Anthropic API that powers Claude. It operates at four points in the workflow, each addressing a distinct operational or research need. At no point does the agent write to the database without staff review. The human judgment call remains with the program manager; the agent reduces the time and cognitive load required to exercise it.
The architecture is additive. The CII Data System as built does not require the agent to function. The agent is a capability layer that sits on top of the existing application, calling the Anthropic API when staff invoke it and populating form fields with draft content for review. The database schema, the scoring form, and the export logic remain unchanged.
The implementation requires one additional API call per agent interaction — a fetch to the Anthropic messages endpoint with the relevant form fields as context. At the Foundation's activation volume, the cost is projected at well under $10 per month. The agent uses the same model (Claude) that generated this document and the CII Data System codebase.
- Program manager enters observed data text for a factor
- Agent reads the text, identifies the applicable rubric tier, and suggests a score with a one-sentence rationale
- Program manager accepts, adjusts, or enters their own score
- Protects inter-rater consistency across staff members scoring different activations
- Highest-value integration — targets the most time-intensive and most variable step in the workflow
- Agent receives all five factor scores, observed data, activation number, and prior activation history
- Drafts the Notes and Recommendations section: identifies weakest factors, proposes specific improvement actions for the next activation
- Program manager edits and approves before saving
- Reduces recommendations drafting from 20 minutes to a 3-minute review
- Ensures recommendations meet specificity standard required by Program Manager scope
- Agent takes a program's full activation history — all CII scores, factor averages, trajectory pattern — and produces a narrative paragraph
- Output is formatted for direct use in municipal briefings, DCEDC presentations, or research partner submission packages
- Converts operational database records into institutional communication without manual drafting
- Consistent framing language across all program reports reduces editorial variation
- Once 8–10 programs have multiple activations scored, the agent analyzes the full dataset
- Surfaces patterns not visible row by row: which disciplines score high on substitution but low on repeat, whether adjacent lift correlates with activation number
- Identifies which proposer or program attributes predict strong pre-commitment campaigns
- This is the research layer — turning the operational database into the evidence base that justifies the strategy's conclusions for external research partners
When the data becomes available.
The work products described in this brief are operational. The following milestones govern when field data collection begins and when the first datasets suitable for analysis will exist.