Building a data strategy for DC’s public charter sector

A five-year strategy to replace fragmented spreadsheet workflows with governed data infrastructure, reusable products, and enterprise intelligence.

When I became Director of Data at the DC Public Charter School Board, the organization had substantial analytical capability but no coherent enterprise approach to data. Important work was getting done, but too much of it depended on spreadsheets passed between teams, one-off analyses, manually maintained reports, and tools that had accumulated without a shared product or architecture strategy. The result was familiar: duplicated effort, inconsistent definitions, avoidable errors, and limited confidence in the agency’s data.

My task was not simply to improve reporting. It was to define how the organization should collect, govern, transform, analyze, and deliver data over the next five years.

The problem

DC PCSB oversees a decentralized public charter sector. Schools submit data to the agency, staff use those data to support oversight and policy decisions, and the public depends on the agency to explain sector performance clearly and credibly.

The underlying coordination problem is difficult:

  • Data arrive through many processes, at different times, and with varying levels of quality.
  • Internal teams need the same information for different purposes.
  • School-facing workflows must be understandable and consistent.
  • High-stakes decisions require traceable and reproducible evidence.
  • Reporting products must serve executives, Board members, schools, researchers, and the public.
  • Legacy systems made it difficult to standardize data models, workflows, permissions, and user experience.

The agency did not need another dashboard or another reporting tool. It needed a strategy for the whole system.

The strategy

I developed a five-year Data Strategy organized around a simple idea: DC PCSB should operate as a product-driven Data Organization.That means treating data infrastructure, analytical workflows, school-facing systems, and reporting products as parts of one enterprise system rather than as separate projects.

The strategy focuses on five connected capabilities:

  1. Reliable data operations
    Improve how data are collected, validated, transformed, and maintained.

  2. Governed data infrastructure
    Establish the warehouse as the authoritative source of truth, with documented access rules, metadata, and lower-risk pathways for broader use.

  3. Enterprise intelligence
    Give staff the ability to ask questions about the DC charter sector and receive timely, trustworthy answers.

  4. Reusable data products
    Replace recurring manual work with standardized products that update when new data arrive.

  5. A product operating model
    Clarify ownership, lifecycle stages, prioritization, user needs, and the responsibilities required to sustain the system.

The strategy is intentionally about choices. It defines what the agency should standardize, what it should stop doing, and where new capacity is required.

From spreadsheets to an enterprise architecture

The core architectural decision was to treat the data warehouse as the agency’s source of truth. Operational systems and school submissions feed governed data pipelines. Those data are transformed into documented, reusable structures. A Data Mart provides a lower-risk, non-PII layer for analysis and reporting. Products then consume the same governed data rather than rebuilding their own versions of the truth.

This architecture is designed to reduce the spreadsheet relay race in which data are repeatedly exported, modified, reinterpreted, and passed from one person to another.

The intended flow is:

submission → validation → warehouse → governed data products → decision

This is not only a technical improvement. It changes accountability. Errors can be traced to a specific stage, definitions can be reused, and products can be updated without reconstructing the entire workflow each year.

Compass: rebuilding the school-facing platform

A major part of the strategy is replacing the agency’s QuickBase-based system, known as the Hub, with a purpose-built web application called Compass. The Hub attempted to solve both school data submission and internal coordination problems, but it lacked a dedicated product strategy and was constrained by the underlying platform. Workflows were difficult to standardize, data models were fragmented, and the user experience reflected the structure of the tool rather than the needs of schools and staff.

Compass is intended to provide:

  • Standardized school-facing tasks
  • Consistent validation workflows
  • Clear status and responsibility tracking
  • Reusable platform components
  • Better integration with the warehouse and reporting systems
  • A product experience designed around actual user needs

In 2026, the agency began using Compass for live school-facing work. The first production workflow focused on LaunchPad configuration validation for the upcoming school year. The immediate objective was modest: prove that schools could complete a real accountability task through the new platform. The larger objective was to establish the foundation for replacing the Hub over time.

Building a reporting portfolio

The strategy also replaces one-off reporting with a portfolio of standard reports. These products are built in Quarto, draw from governed Data Mart data, and are designed to update as new data become available. Rather than asking analysts to recreate the same work every year, the agency can invest once in a durable reporting product and improve it over time.

The emerging portfolio includes products such as:

  • DC CAPE assessment reporting
  • Sector-level ASPIRE reporting
  • Public performance reports
  • School finance reporting
  • Internal decision-support products

Compass will eventually include a Reporting Center where users can access reports based on their permissions. This creates a single delivery channel while allowing the agency to understand who is using each product and where additional support is needed.

The shift is from producing documents to managing products.

Enterprise intelligence

One of the strategy’s central goals is to build Enterprise Intelligence: the ability to ask a computer a question about the DC charter sector and receive a trustworthy answer. That requires more than visualization software. It depends on:

  • Governed and well-modeled data
  • Shared definitions
  • Traceable calculations
  • Reusable analytical code
  • Clear documentation
  • Appropriate tools for different users
  • Staff who understand how to interpret the results

Tableau can help staff quickly create high-quality visualizations connected to the Data Mart. Quarto can provide standardized analytical reports. Compass can deliver those products to the right audiences. None of those tools is sufficient alone. Enterprise Intelligence is the capability created when they operate as one system.

Changing how the Data Team works

The strategy also required a change in operating model. I introduced or strengthened practices including:

  • Product ownership and explicit lifecycle stages
  • Two-week iterations managed through GitLab
  • Ranked portfolio prioritization
  • Capacity tracking and sprint integrity monitoring
  • Public documentation through a Data Team Handbook
  • Standardized project initiation for cross-functional work
  • A clearer distinction between operational requests, analytical projects, and durable products
  • Greater emphasis on reproducibility, code review, and institutional memory

These practices are designed to protect the team from becoming a permanent queue of disconnected requests. They also make tradeoffs visible. When the organization asks for more work than the team can complete, the answer should be a prioritization decision—not hidden overwork or lower quality.

Progress so far

The strategy is still in its first year (as of writing this in July 2027). Most of the work is therefore foundational.

Progress to date includes:

  • Adoption of the five-year Data Strategy as the organizing framework for the Data Team
  • Continued development of the warehouse and Data Mart
  • Launch of the first production Compass workflow
  • Definition of a reporting portfolio and Compass Reporting Center concept
  • Development of standardized Quarto reporting products
  • Increased use of GitLab for portfolio and product management
  • Explicit governance work around access, documentation, and data use
  • A reset of the agency’s Enterprise Intelligence approach
  • Greater executive engagement with data architecture, product strategy, and technical transparency

Some parts of the strategy are ahead of the original schedule. Others remain early. The principal constraint is not the absence of ideas; it is the managerial and technical capacity required to replace legacy workflows while continuing to deliver current-year oversight work.

What I have learned

A data strategy is not a technology roadmap. It is a set of choices about how an organization will make decisions, assign responsibility, manage risk, and build trust. The hardest part is rarely selecting a database or reporting tool. The hardest part is stopping fragmented practices that feel locally convenient but create enterprise-wide costs. The work has also reinforced several lessons:

  • A source of truth only matters if products actually use it.
  • Standardization is a product decision, not merely a data-engineering task.
  • Transparency improves quality only when teams are willing to expose unfinished or imperfect work.
  • New platforms fail without product management, regardless of their technical architecture.
  • Reporting should be treated as a maintained portfolio, not a collection of annual deliverables.
  • Strategy requires saying no to work that does not support the chosen direction.

What comes next

The next phase is execution at greater scale. Priorities include:

  • Expanding Compass beyond its initial workflows
  • Delivering the first mature set of standardized reports
  • Establishing the Reporting Center
  • Improving Data Mart usability and metadata
  • Strengthening governance and documented access pathways
  • Building staff capability to use live, governed data
  • Increasing managerial and product capacity
  • Publishing selected code and documentation where transparency can improve trust and external review

The long-term objective is straightforward:

DC PCSB should be able to collect data once, govern them consistently, reuse them across products, and answer important questions without rebuilding the system each time.