In-Depth

DCStat

I. Project Information

Project Designation: DCStat

DCStat is part of the District of Columbia’s broad management strategy to increase open governance improve citizen services. DCStat aims to help government agencies operate as leaner, more responsive, higher performance organizations that fulfill their responsibilities to residents. DCStat addresses various problems including inefficient agency operations, inability to share and use data across agencies and inability to deliver services proactively.

DCStat Objectives

DCStat was established as a set of analytical processes and data-driven application tools to support Mayor Anthony A. Williams’ Hot Spot crime reduction initiative and raise the quality of life for District residents through better neighborhood safety and public health. As the capability has grown, other initiatives have been supported by DCStat analytics, including New Communities, Nuisance Properties, Operation Fight-Back and Clean City.


The most significant analytical process within DCStat is the City Administrator’s (CA) Executive Briefing program, which aims to enhance public safety, improve service delivery, reduce costs and increase revenues in the District. These briefings provide a regular forum for District leaders to gain operational transparency into agencies, evaluate performance and hold managers accountable. The briefings also provide a dynamic environment where the CA can reinforce his vision of a culture of continuous improvement, encouraging agencies to adapt and innovate.

DCStat Business Risks

Program managers and officials faced two business risks when implementing DCStat. First, agencies were reluctant to cooperate with developers and share data with other officials and agencies. Prior to the implementation of DCStat, agencies collected and used data for internal purposes. DCStat required opening up databases and files to outside scrutiny.


Agency employees also expressed concern that the new automated DCStat system would replace jobs. The automated system promised to make operations, such as data analysis and record-keeping, more efficient and would in turn make manual labor for those jobs obsolete.

User Impact of DCStat

DCStat serves four different constituencies and consists of five application “views”: (1) Neighborhood View (2) Mobile View (3) Analyst View (4) Executive View and (5) Residents View. Each view requires varying levels of training and has different levels of interactivity, offering different reporting, mapping, query and analysis functions.

Neighborhood View: Supporting District Staff with In-Depth Information

The Neighborhood View is DCStat’s web-based application intended for broad use by District staff. Neighborhood View supports activities of the Mayor’s Office of Neighborhood Services for strategic initiatives, including Hot Spots and New Communities. At their desk and during public meetings, Neighborhood Services Coordinators may access up-to-date information regarding city services, crimes and other activities in specific areas of the city. They may produce maps and graphs to analyze incidents, compare geographies and discover trends.

Mobile View: Facilitating Field Workers’ Fulfillment of Service Requests

The Mobile View provides access to DCStat using wireless devices. From a PDA, mobile workers query DCStat using site address, service numbers, property owners’ names and other criteria. Mobile workers can obtain a property summary report that identifies owner and tax assessment information, service request history, housing violation citation records and crime activity associated with a property. Details are available through drill-down screens. The system will locate similar nearby incidents to help identify cause or patterns.

Analyst View: Experts Examining City Information

The Analyst View will be a secured or private view, available only through special access given by the DCStat administration. It will require a higher level of training and interactivity and will allow for expert dashboard analysis and reporting functions.

Executive View: Empowering District Officials

The Executive View will be available via the intranet and will require minimal training. Municipal officials will use Executive View to perform executive data analytics once they receive an alert, watch notification or simply access DCStat and want more information on a scenario. This version also will offer reporting functions.

Residents View: Realizing Residential Services…in Minutes Rather Than Months

The Residents View is a public version that will be available via the Internet. It will require no training and will allow for basic interactivity that permits residents to perform neighborhood comparisons and view what is happening in their location. This view also will allow residents to perform general mapping and query functions.


II. Organizational Objectives

Short and Long Term Benefits of DCStat

District of Columbia officials and managers developed DCStat to make municipal government more efficient, effective, transparent and accountable. To date, DCStat has realized these four goals in the following ways. DCStat:


Provides Data Coordination and Real-Time Data Retrieval

Previous systems only allowed for limited data analysis because of paper-processes, disparate databases and the availability of outdated data. DCStat automatically refreshes data from various city agencies on a regular and frequent basis. Even better, DCStat’s ability to rapidly gather, track and analyze data allows city officials to make decisions based on real-time, dynamic information and current conditions. This reduces wasteful spending and increases revenues, enhances public safety and improves service delivery.


Because DCStat continually monitors performance and service delivery, city managers can pinpoint when and where municipal services are needed the most. DCStat’s GIS mapping capabilities show city officials, via desktop or personal handheld devices, the Hot Spots or troubled areas that need immediate attention. The ability to link data and automatically identify areas in need allows for faster and better decision-making, which dramatically improves the quality of life and business conditions in the District of Columbia.


Increases Government Efficiency

DCStat improves government operations by: (1) Providing access to real-time information; (2) Creating agency specific applications; (3) Improving decision making by providing more data of higher quality; (4) Accelerating data sharing between agencies; (5) Presenting data to the City Administrator in a format that shows agency performance; (6) Increasing transparency within agencies; and (7) Supporting all major initiatives of the City Administrator.


Forges Fiscal Benefits

DCStat also will provide fiscal benefits including:

  • Higher performance due to heightened transparency and measurable performance goals
  • Faster service delivery times due to residents’ ability to track and follow fulfillment of service requests
  • Reduced costs and increased revenue due to efficient allocation of resources, elimination of duplicated efforts, balanced workloads to increase service quality.

Within its first year of operation, DCStat has already achieved over three times its return on investment. The most recent Executive Briefing session identified nearly $7 million in uncollected revenue by analyzing one of the 62 city agencies in the District. Additional cost savings through recommendations to reduce waste and improve performance add an estimated $4 million to the bottom line. In addition to uncollected revenue, DCStat provides potential cost savings by realigning business processes within the agencies and fostering coordination between agencies.


Improves the Citizens Service Request Center (SRC), increasing government transparency and accountability

The District’s dynamic Service Request Center (SRC) offers real-time status of every request, 24 hours a day, seven days a week. Residents will be able to access their service request from the Internet using DCStat’s Residents View to place a service request or view the status of a request. Agency employees can access Neighborhood View to map the location of residents’ service requests. The new technology allows agencies to coordinate their efforts to efficiently and effectively respond to the residents’ needs. DCStat shows real-time status at any time, and offers complete tracking of service delivery and agency performance.


Debunks Data Storage for Efficiency and Effectiveness

Prior to the implementation of DCStat, many city agencies used paper-based processes or outdated, inaccessible databases. Disconnected databases across the city stored data in aging legacy systems. Because databases were not linked or accessible to other agencies—and often times even inaccessible to different departments within an agency—information was inaccurate and incomplete.


Develops Interagency Cooperation and Data Access

DCStat allows agencies and departments to share and leverage others’ data, making the allocation of resources more efficient and effective. Managers now can access performance results as well as information about each department and employee. This cross-department and cross-agency data sharing heightens managers’ awareness of the "big picture," making managers more accountable. Data also will be accessible in five views for different audiences depending on the viewers’ level of training and depth of analysis he or she needs. The user community will include: District residents, mobile workers, agency workers, agency and District leaders, and analysts, using Residents View, Mobile View, Neighborhood View, Executive View and Analysts View, respectively.


Leads the Nation’s Cities Through Transforming Government

DCStat is transforming the way municipal government and city services work by bringing private sector efficiencies to the public sector and establishing the District of Columbia as a leader in municipal performance, accessibility and functionality. The unprecedented transparency and accountability driven by DCStat provides a more responsive and effective city government for District residents and businesses.

b) How does DCStat Work? DCStat’s Purpose

DCStat aims to help government agencies operate as leaner, more responsive, higher performance organizations that fulfill their responsibilities to residents. DCStat addresses various problems including inefficient agency operations, inability to share and use data across agencies and inability to deliver services proactively.


DCStat was created to support Mayor Anthony A. Williams’ Hot Spot crime reduction initiative and raise the quality of life through improved public health and neighborhood safety. DCStat also is the primary tool supporting the City Administrator’s (CA) Executive Briefing program.

c) How does DCStat Work? DCStat’s Features

The key features of DCStat are in the areas of data and applications. On the data side, DCStat provides access to near-real-time data from the various DC government agencies that have been integrated in the DCStat data environment. On the application side, key features include the ability to search and analyst available agency data and to view results spatially in map form and other logical groupings in report form.

d) How does DCStat Work? DCStat’s Functions

DCStat utilizes EAI (Enterprise Application Integration) technology to interface agency data into the DCStat data environment on a near-real-time basis. The value to users is up-to-date data on which to base timely decisions and the ability to correlate available agency information for more coordinated actions in delivering services. DCStat application functions include data searching, mapping, watches and address summary.

  • Data searching enables users to query and analyze data within the data repository based on type (address, service request, property owner), data range or text string.
  • Mapping allows users to select data sets for display on a map, perform map actions (i.e. identify, search, zoom) and gives users the ability to view detailed data records.
  • Watches enable users to define monitors of conditions around the District (i.e. number of homicides, number of potholes) and benefit from the near-real-time data interfaces that trigger notification when threshold levels are reached.
  • Address summary allows a user to enter an address and obtain a summary of information including the property owner’s name, open/closed service requests, open/closed inspections, open/closed complaints and other property status information.
e) DCStat Champions and Challengers

The District initially developed DCStat to support Mayor Anthony A. Williams’ Hot Spot crime reduction initiative but its uses and applications have grown significantly since idea inception. Most notably, DCStat inspired a cultural shift from one of agencies acting in isolation to a new era of cooperation and cross-agency data sharing.


For example, agency leaders for the District’s Department of Consumer and Regulatory Affairs (DCRA) were eager to change their antiquated systems to DCStat, but some employees expressed concern about implementation. Employees fear about automated systems replacing jobs and concern about the system exposing fraud or abuse made some District employees reluctant to accept DCStat. However, employees soon began to see the benefits and efficiencies of working with an automated system and became more accepting of DCStat.


f) User Feedback Proved Critical to DCStat’s Development

During the development of DCStat, software engineers and project managers worked with authorized users in District agencies to obtain feedback about ways to improve the system. District officials and agencies set requirements or cases to measure the progress of development and also participated in regular and frequent user acceptance testing. User testing throughout the build process helped DCStat developers enhance the application from user feedback suggesting improvements.

g) Organizational Challenges in Completing DCStat

Many District workers have operational databases in desktop application format not linked to sequential tasks. DCStat had to find a way to incorporate this information. Furthermore, the GIS technologies are not industry standards, making it difficult to modify or build on top of them because for the most part you cannot customize the software. DCStat also had to work against some individual’s parochial views of data exchange. Though data sharing actually enhances agency performance, not everyone was excited to open their operations to external view.


In addition, navigating the sea of proposals, budgetary constraints and a multi-tiered approval process is a challenge unique to public sector development projects. The DCStat team ably worked through this by educating the user community about the government process and communicating with users about the development time line in order to shift expectations.

h) Accomplishment of Goals

Consistent with Mayor Anthony A. Williams’ original goals, DCStat today incorporates data including crime statistics but also gives users access to a vast range of District information on citizen services. With each version of the four different views, developers incorporate user feedback to improve the system according to user needs.


As the project progressed, developers integrated more data sets from the District agencies and outside sources, such as Dun and Bradstreet business information and Federal points of interest. As users provided feedback to developers, the scope of DCStat grew from providing city leaders and managers with a management tool to help reduce crime to a program of significant use to different audiences including mobile workers and District residents.


III. Category

DCStat is proud to submit this application for consideration in the E-Business Application Development category of the 2006 Application Development Trends Innovator Awards.


IV. Methodology/ Process

DCStat’s Project Management Methodology

DCStat’s project management methodology includes the normal phases of plan, design, build, test, deploy; however, design-build-test are an integrated phase that aligns with the agile software development approach.

Agile Software Development Life-Cycle Methodology

DCStat employs an agile development approach that focuses on "use case" requirements, practices test-driven development, schedules shorter fixed-time iteration development cycles and engages stakeholders consistently throughout the SDLC. To ensure timely and high-quality results, the development process is formally tracked through daily stand-up meetings and an automated build process that facilitates integration, regression and deployment of application on components on a continuous basis.

Stakeholder Reviews

Throughout the integrated design-build-test phase of development, DCStat stakeholders validate requirements and perform hands-on reviews of features as the applications are being designed and built. For instance, over the last two-week development cycle, stakeholders were engaged at the onset to define and confirm requirements. Then stakeholders also attended design review sessions every two weeks to perform user acceptance testing (UAT), provide feedback on application functionality/interfaces and sign-off on the finalized application. The results are applications that better support the stakeholders’ business analytical needs.

Implementation Timeline Summary

The implementation of DCStat will take place in three phases. Currently, DCStat is in the second phase. Each phase has a design, data content and view development stage.


Phase I: Prototype

The first phase of DCStat was a Prototype or Pilot Program launched with seed money in March 2004. The City Administrator observed the first functional prototype in April 2004. Shortly thereafter in May 2004, OCTO set up the first live feed from the Mayor’s Call Center database to DCStat.


Phase II: Formal Funding and Build

Phase two began in July 2004 when DCStat received funding to develop its program. Neighborhood View 1.0 was released in November 2004 and completed full infrastructure build-out in January 2005. Other phase two accomplishments include: the foundational architecture for the enterprise data repository, where all agency data will reside, Mobile View 1.0 release and Executive Briefing.


Phase III: Fine Tuning the Infrastructure

The third phase will include infrastructure tuning, increased data repository capabilities, and other application releases. On the data repository side, we will expand the data catalog, integrate and publish additional agency data sets, and develop business intelligence analytical and reporting functionality. On the application side, this includes Analyst View, Executive View, and Residents View, Neighborhood Services Action Plan, collection/alert/notification engine, enhancements to Neighborhood View and Mobile View.

e) Development Tools

DCStat uses a variety of development tools to support the agile SDLC process. These tools fall into the categories of Integrated Development Environment (IDE), Core Development Tools, Testing Tools and Software Developer Kits (SDKs), and will be elaborated on further.

Integrated Development Environments (IDE) Tools

The base technical environments that developers on the DCStat team are using include Eclipse, Visual Studio, Rational Application Developer (RAD) and to a lesser extent, DB Visualizer for database related development efforts. These IDE tools provide an application framework and development platform for developing software applications.

Core Development Tools

DCStat has used an extensive set of tools to support development efforts, above and beyond the IDE platforms. These tools include Cruise Control which serves as the continuous integration framework, Subversion which provides source control of component code, Ant/NAnt/Maven that supports auto-build, Mantis which enables issue/bug tracking and task management, XML Spy for XML modeling, editing, debugging, and transform, and miscellaneous open source tools for documentation (JavaDoc, XDoc) and logging (Log4J, Chainsaw).


Of all the core development tools, Subversion, Cruise Ctrl and Anr/NAnt/Maven support the critical development function of auto-build that is the basis of DCStat’s agile SDLC. The auto-build function provides integration, regression and status of components on a continuous basis. Nightly auto-build status is reviewed at daily developer stand-up meetings to determine if any components "broke-the-build" and the appropriate course of action for the team to resolve any issues.

Testing Tools

Given the significance of testing within any IT/IS capability, the DCStat program has incorporated numerous testing practices and tools to ensure quality products. On the practices side, testing is integrated into the development with developers required to design and build test components before they even begin designing and building functional components. The value in doing this is that developers identify potential failure points with the functional component before coding of these components even starts, which saves time and enhances quality. The test components that are built are used individually by developers and also incorporated into the auto-build process that tests integration and regression Testing tools used on the DCStat program include N-unit, J-unit, HTTP-unit, Grinder and J-Meter. Testing tools to be used in the near future will include tools from the Mercury Interactive suite for testing end-to-end application functionality and potential stress/performance issues.

Software Developer Kits (SDKs)

The final set of "tools" that developers will utilize through the SDLC will include SDKs, which are the toolkits for integrating various custom and packaged software components within the overall application architecture. SDKs identified include the Java SDK which is a java development tool, the ESRI SDK for GIS integration, the Business Objects SDK for business intelligence integration and potentially the HP Openview SDK for integrated operations management.


V. Technology

Technical Challenges

The greatest challenge developers faced when building DCStat was integrating with a diverse set of agency legacy systems to obtain data. These systems ranged from structured application/database platforms to unstructured productivity based templates, such as Word, Access and Excel.

Software Tools

DCStat is an n-tier application built on an open Services Oriented Architecture (SOA) and compliant with industry standards for interoperability and portability. It is built on the J2EE platform for backend functionality and services, with .NET as the presentation on the front-end. It makes extensive use of Web Services, SOAP, XML and DHTML, along with Oracle, Oracle Spatial, ESRI ArcSDE, and customized GIS-enabled, Business Intelligence modules and knowledge discovery. It uses SeeBeyond's e-gate for moving data from various data sources to the DCStat Enterprise data warehouse.

Software Selection

DCStat developers used a variety of development tools discussed above. The rationale for tool selection was: (1) open source software that proved reliability, scalability and stability; (2) cost-effective products; and (3) familiarity to the development team members. Going forward, DCStat will adopt/ migrate tools recently selected tools at the enterprise level.

Service Oriented Architecture

To achieve the SOA infrastructure for DCStat, the District of Columbia implemented Sonic ESB (enterprise service bus). The District selected Sonic Software for this requirement because Sonic ESB was suited to address the need for publishing and consuming distributed application services. It also allows the flexibility to integrate service providers in the future with no additional customization. Sonic ESB also provides a Continuous Availability Architecture (CAA), ensuring that communications and interactions between services are always available.


Other technologies used within DCStat that will be integrated into the SOA include a robust search engine, alerts/notifications, Geographic Information System (GIS), Business Intelligence and Oracle database technologies.
Delivery Network: ESB achieves reliable messaging transport between services.


Interface: DCStat publishes shared standalone services using interoperable web services.


Governance Semantics: An Information Model bridges service and data islands by adopting a common abstract reference model and document-style XML.

Key Infrastructure Features

Other key features of the infrastructure focus on the delivery of data and providing users with an integrated view. For instance, users can see data in GIS or report format through Business Objects but all data will eventually feed into a portal accessible to District officials, managers, employees and residents.

What’s Next?

As DCStat evolves from capital phases to an operational project, it will comply with a suite of standard products widely used in OCTO. While DCStat uses some of the OCTO standard products, such as Dimensions, Cruise Control and Mercury Interactive, it has not yet worked with OCTO’s enterprise architects to aggressively use only standard products. The continuous build feature of DCStat allows developers to begin replacing open source software, including CVS, Subversion, Mantis, Share Point, Visual Studio and Eclipse with licensed products. DCStat currently is installing tools such as IBM Websphere and Radical Application Development (RAD) to replace the open source software in order to comply with OCTO standards.


VI. Project Team

Development Team Size

21 individuals participate on the development team, 12 of whom are developers.

Development Experience of Team Members

The developers are .net, J2EE project managers and general application architects as well as high-level enterprise application integration (EAI) architects.

Professional Development

DCStat provided software training for developers through its vendors including FastSearch and Sonic. Other applications, such as Websphere, have computer based training (CBT) programs to support developers.

Calendar Timeframe

Initially, the DCStat project development team created a formal timeline for creation and a phased rollout of the five views and advanced versions. Since DCStat began in March 2004, the DCStat team has adjusted the timeline accordingly to accommodate shifting demands and expectations.

Assessment

Review of the DCStat program indicates success based on the Executive Briefing process and agency user feedback. DCRA Commissioner Patrick Canavan stated, "DCStat has provided a structured review process, business analysts, tools, and data to supporting our agency’s strategic objectives." DCStat supported the Executive Briefing in the following specific ways:

  • DCStat analysts assessed DCRA business operations and data to identify and prioritize 16 of the agency’s issues. DCStat reported on the top five issues and presented key findings.
  • Based on these key findings, management at DCRA and related agencies worked together to create a strategy for improvement.
  • DCStat provided an objective validation of issues by allowing DCRA to see where rental properties are located, whether they are in compliance with licensing and inspection, where services are requested and where inspections are taking place, and highlighted persistent problems.
  • DCStat has also helped us identify the need to ensure properties are classified properly so the Office of Tax and Revenue can apply the appropriate tax rate.
  • DCStat provided DCRA employees with Neighborhood View, an analytical tool with strategic agency datasets, for referencing information about inspections, ownership, property, tax, and other related data. Neighborhood View has reduced the time and effort spent through the process of investigating properties and their owners. Additionally, Neighborhood View allows employees to view data on maps, setup watches to monitor when predefined conditions and threshold levels are reached, and drilldown into the details of individual data records.
Review of DCStat: Looking Back

One change DCStat developers would make if they had to do the project over again is to build in elements to the development strategy that would not have required the DCStat prototype to become the base application. Developers were required to build a prototype under a tight deadline of just 60 days, and project timelines forced developers to base the DCStat application off of the prototype. Developers instead would have preferred greater flexibility in the prototype and application building cycles.