Work Products Overview — Duncanville Arts Foundation
Cultural Investment Strategy — Infrastructure Overview

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.

Foundation
Duncanville Arts Foundation
Strategy Document
Cultural Investment Strategy v2.0
Prepared For
Research Partners
Date
February 2026

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.

The Research Question

"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 data gap this work fills. Southwest Dallas County has no prior structured evidence base for cultural investment outcomes. National studies exist. Duncanville has none of its own. This is not a limitation of the strategy — it is the reason the strategy was designed the way it was: to generate local, reproducible, behaviorally grounded data from the first activation forward. The Foundation is not asking a research partner to validate an existing body of evidence. It is asking one to help build the first.

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.

Primary URL
duncanvillearts.org
Platform
Squarespace
Custom HTML injected via Code Blocks
Intake Method
Formspree
Form data forwarded to Foundation email
Primary Audiences
4
Proposers, municipal, economic dev, research

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.

Website Structure — Section Sequence
01 — Hero Cultural Investment Strategy Economic framework & research mandate 02 — Economic Basis The $31M Framework BLS + Census derivation. Cited methodology. 03 — The Work Research Mandate Two-path outcome structure. Evidence-first framing. 04 — Pipeline Five-Stage Process Intake → Develop → Validate → Activate → Graduate 05 — CII Scoring Cultural Investment Index Full formula, five-factor rubrics, graduation thresholds 06 — Arts Junction Old Rail Station Venue Testing venue, property partnership, activation calendar 07 — Propose Intake Portal Nine-section form → Formspree → Foundation email Bibliography & Footer 6 cited sources. BLS, Census, ORS property data, CIS v2.0, derived calculations. All verified Feb 2026.
Research relevance. The website's bibliography cites all six sources underlying the $31M expenditure derivation, including the specific BLS Consumer Expenditure Survey table, the ACS household data, and the CIS itself. The derivation methodology is reproducible and documented in the public domain. The site functions as a reference document for understanding the data inputs that precede field measurement.
01Economic baseline section presents the $31M derivation with source citations and a conditional recapture scenario stated as projection range, not guaranteed outcome.
02Pipeline section communicates all five stages with validation gate language. The 100% pre-commitment threshold is stated as a governing principle with its rationale.
03CII Scoring section reproduces the complete Appendix D methodology including the composite formula, all five factor weights, four-tier rubrics, and the multi-activation graduation framework.
04Proposer intake routes through a nine-section HTML form capturing all Appendix A fields. Submissions are forwarded to Foundation email via Formspree. A confirmation page and acknowledgment email complete the intake loop.
05Bibliography lists all six primary sources with access dates, all verified February 2026. Claims on the site are traceable to their source in a single citation.

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.

Connection to research validity. Execution consistency is a prerequisite for data comparability across activations. The Program Manager scope standardizes three things that directly affect CII scores: the timing of data collection, the method of survey administration, and the procedure for calculating adjacent business lift. Without this standardization, scores from different activations would not be methodologically comparable — a requirement for multi-activation trajectory analysis (CIS v2.0, Appendix D, Section 5).
Replication relevance. If the Foundation's model is later replicated in other Southwest Dallas County municipalities, the Program Manager scope provides the execution template that makes replication faithful to the original methodology. The role definition travels with the strategy.
Phase 01 — Pre-Activation
Validation Campaign Management
  • 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
Phase 02 — Activation Day
Execution and Live Data Collection
  • 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
Phase 03 — Post-Activation
Data Processing and Reporting
  • 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.

System Architecture
User
Program Manager
2–5 users, individual credentials
User
Program Manager
Any device, any browser
User
Foundation Director
Dashboard + export access
↓ HTTPS encrypted connection
Front End
cii_app.html — Hosted on Netlify
Login • Scoring form • Dashboard • Reports • Export
↓ Supabase JavaScript client library
Supabase Service 01
Authentication
Email + password login. Row-level security — only authenticated users access data.
Supabase Service 02
PostgreSQL Database
Two tables. Standard SQL. Full export at any time.
Data Outputs
Export 01
All Activations CSV
Full dataset — all factors, all observations, all scores
Export 02
Programs Summary CSV
One row per program — avg CII, latest score, determination
Export 03
Multi-Activation CSV
CII trajectory per program across all activations
CII  =  (Pre-Commitment × 0.30) + (Resident Share × 0.25) + (Substitution × 0.20) + (Repeat × 0.15) + (Adjacent Lift × 0.10)
F1
Pre-Commitment AchievementTicketing platform, sponsor docs, vendor deposits
× 0.30
F2
Duncanville Resident ShareSurvey Q1 ZIP codes + ticketing platform (ZIPs 75116, 75137, 75138)
× 0.25
F3
Substitution SignalSurvey Q2/Q3 — % who would have spent outside Duncanville
× 0.20
F4
Repeat ParticipationSurvey Q5 + ticketing repeat-purchase data — default 50 for first activations
× 0.15
F5
Adjacent Business LiftORS tenant POS or foot traffic vs. 4-night non-activation baseline
× 0.10
programs One row per program
iduuid / primary key
titletext
proposertext
disciplinetext
pipeline_stagetext
statustext
notestext
created_attimestamptz
activations One row per scored activation
iduuid / primary key
program_iduuid → programs
activation_date / numberdate / int
total_attendanceinteger
survey_responses / rateint / numeric
f1_score through f5_scorenumeric (0–100)
f1_weighted through f5_weightednumeric
f1_obs through f5_obstext (observed data)
cii_totalnumeric (0–100)
determinationtext
strengths / development_areastext
recommendations / trajectory_notetext
Tables are linked by program_id — every activation record knows which program it belongs to. Row-level security ensures only authenticated users can read or write. The database is standard PostgreSQL and exports as a full SQL dump at any time.
01 — Within 24 hrs
Raw Data Collected
Ticketing report, survey responses, ORS tenant POS data pulled and compiled by Program Manager.
02 — Within 7 days
CII Scored
Program Manager logs into CII Data System. Enters observed data and factor scores. System calculates CII composite and determination automatically.
03 — Within 14 days
Proposer Feedback
Scored worksheet delivered to proposer with Notes and Recommendations completed. Development support or sunset conversation initiated if applicable.
04 — Reporting cycle
Data Exported
Foundation exports All Activations CSV from the Data System for external research reporting. All five factor scores, observed data, and CII totals included.
Research data availability. Every activation the Foundation scores produces a complete record: five factor scores with supporting observed data text, the composite CII, and the determination. The multi-activation export provides a CII trajectory for every program across all activations — the primary dataset for pattern analysis at the program level. The full activations export provides the complete row-level dataset for cross-program and cross-discipline analysis. Both are available in CSV format from the Data System immediately upon export.

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.

Agent Integration Architecture
Staff Action
Program Manager Invokes Agent
"Draft Recommendations" or "Suggest Score" button in CII Data System
↓ Form fields sent as context
Context Sent
Observed Data + Scores
Factor observations, activation number, program history, CIS rubric criteria
API Call
Anthropic API
claude-sonnet model. Response returned in under 3 seconds.
↓ Draft content returned to form field
Staff Review
Program Manager Accepts, Edits, or Overrides
No agent output enters the database without staff approval
↓ Staff saves record
Database
PostgreSQL — Supabase
Record saved with staff-reviewed content. Agent involvement is not stored separately.
Integration 01 — Scoring Assistance
Factor Score Suggestion
  • 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
Integration 02 — Recommendations Drafting
Post-Activation Development Guidance
  • 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
Integration 03 — Narrative Report Generation
Municipal and Research Reporting
  • 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
Integration 04 — Cross-Program Pattern Analysis
Research-Layer Insights
  • 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
01 — Staff enters data
Observed Data Field
Program manager types: "100% reached 5 days before deadline. Two revenue sources. No extensions."
02 — Staff clicks button
"Suggest Score"
The observed data and CIS rubric criteria for Factor 1 are sent to the Anthropic API as context.
03 — Agent responds
Draft Score + Rationale
Agent returns: "Suggested score: 85. Strong tier. Commitment reached before deadline with multiple revenue sources and no extensions."
04 — Staff decides
Accept or Override
Program manager accepts 85, adjusts to 82 based on additional context, or enters their own score. Record saves with the staff-approved value.
Research integrity note. The agent assists with scoring and drafting but does not determine outcomes. The CII score that enters the database is always a staff-reviewed value. This preserves the methodological integrity of the dataset for research purposes — the scores reflect human judgment informed by the CIS rubrics, with the agent functioning as a consistency reference rather than an autonomous rater.
Development status. The AI Agent layer is planned for development following the completion of Arts Junction intake and the first activation cycle. The architecture is defined. The CII Data System is built to receive it without structural changes. Estimated development time once initiated: two to four weeks.
01Technical dependency: Anthropic API key. No additional infrastructure required beyond what the CII Data System already uses.
02Projected cost: Under $10 per month at the Foundation's activation volume. Fractions of a cent per scoring worksheet interaction.
03Staff oversight: Every agent output is a draft. No agent content enters the database without explicit staff approval. The human decision is preserved at every step.
04Research value: Consistent scoring language and standardized recommendations framing across all activations improves the comparability of qualitative fields in the dataset — a direct benefit to external research analysis.
05Model: Claude (Anthropic). The same model used to develop the CII Data System codebase and this document.

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.

Q1 Certificate of Occupancy — pending. The City of Duncanville has not yet executed the CO for Arts Junction at Old Rail Station. Formal intake opens upon receipt. No target date is confirmed; the Foundation is in active contact with the City.
Q1–Q2 First intake cohort. Immediately following CO execution, the Foundation opens intake and begins processing the first proposer cohort through the five-stage pipeline. Development workshops begin. Validation campaigns launch.
Q2 First scored activations. Programs that achieve 100% pre-commitment activate at Arts Junction. The first CII Scoring Worksheets are completed. The first activation records enter the CII Data System. Field data collection begins.
Q3–Q4 Multi-activation data available. Programs reaching their second and third activations produce the trajectory data required for graduation decisions and pattern analysis. This is the first dataset suitable for cross-program analysis.
Year 1 Full Year One dataset. Up to 16 activations across six artistic disciplines, scored across all five CII factors with complete observed data, survey results, and adjacent business measurements. The first structured evidence base for cultural investment outcomes in Southwest Dallas County.