Continuums of Care (CoC) are searching for answers.

There is an ever-increasing necessity to target limited resources and use data to make funding choices.

Communities need to access data to analyze their homeless systems and make thoughtful, data-driven decisions about policy, programs, and funding.

HMIS are complex and require dedicated staff who can manage, extract, and analyze data. That data needs to be transformed into easily consumable information for decision-making. Resources are scarce.

Many CoCs are left scratching their heads – How do we overcome these challenges? How do we start building a data culture? Snohomish County Human Services Department (HSD), the CoC for Snohomish County, Washington might have one answer. They created data visualizations, including public-facing data dashboards, as part of a bigger plan to create a performance culture and transform their homeless system.

Example of a performance dashboard that provides valuable at-a-glance information on core measurements of homelessness
Infographic showing data from HMIS on exits to permanent housing from emergency shelter and rapid re-housing
Snohomish County HMIS Team: Stephanie Patterson, Alex Vallandry, Jesse Jorstad and Kendall Shawhan

Creating a Community-Wide Dashboard

In 2009, Pierce, King, and Snohomish Counties began participation in a three-county initiative funded by the Bill & Melinda Gates Foundation in partnership with Building Changes, a local non-profit, aimed at building cross-sector partnerships to improve coordination and access to services for families and youth experiencing homelessness. Adopting a data-driven culture is a key strategy of the initiative. To track progress on these goals, Jesse Jorstad, Lead Data & Program Analyst for HMIS with Snohomish County Human Services Department (HSD), spearheaded development of a community-wide dashboard to support and improve data analytics. Since that time, the Snohomish County HMIS team has enhanced the dashboards to include HUD’s System Performance Measures (SPMs), ad-hoc reporting capabilities, developed data analysis tools with more sophisticated data presentations, and leveraged these new data tools to launch a data-driven culture initiative.

The Snohomish County HMIS team has been a driving force behind use of these dashboards to transform the system and build an understanding of the data landscape. The team undertook a community-driven development process, seeking continuous input and refinements from key partners. Supported by Snohomish County leadership, the team built trust in data by using it to answer questions and teaching community members how to interpret complex data. This iterative process transformed the basic data dashboards into near real-time data analysis tools that have been central to more effectively using local data to make decisions about policies and practices as well as troubleshoot system challenges. The dashboards have fundamentally changed the way that HMIS data are used in Snohomish.

"Initially I had this vision that I was going to develop a handful of dashboards and at some point, I was just going to hit the refresh button every month. In reality, it is an iterative process that I work on every day to answer the next question and make the data more visually apparent. It’s always asking, 'why does the data look like this' and meeting with the appropriate groups to find those answers." ~Jesse Jorstad, Lead Data & Program Analyst for HMIS, Snohomish County Human Services.
The Snohomish County HMIS Team Meeting About Data

Data, Performance, and Policy - Understanding the System

Data visualization dashboards are now being used to make major decisions about housing strategies in Snohomish County.

Built with the ability to drill into the data in near real-time as questions arise, the dashboards have led to a better understanding of family homelessness. To assist the team focused on family homelessness, the Snohomish County HMIS team built a “data story”, using a combination of published and ad-hoc dashboards that provide understandable information to support specific policy discussions. As an example, these dashboards illustrated how families in rapid re-housing programs experienced shorter lengths of time homeless, fewer returns to homelessness, and higher rates of income progression – resulting in better housing outcomes for families. At the same time, the dashboards showed that transitional housing was lagging behind on two measures: average length of stay and income progression. Based on this information, the CoC decided to narrow the populations being served by transitional housing and target resources to the expansion of rapid re-housing. Coupled with a new policy to link emergency shelters and rapid re-housing projects so that clients in shelters have better access to rapid re-housing projects, Snohomish County was able to reduce the length of time spent in shelters by families from 278 days to 97 days.

Having interactive dashboards to facilitate understanding of the system from multiple angles when talking about policy allows the CoC to make broad policy decisions more quickly because questions can be answered immediately.

Sample dashboard showing households served in rapid re-housing programs between 2016 and 2017
Example rapid re-housing household dashboard showing exit success rates, # of enrollments, days to housing, program exits, and average length of stay in project types
"We used to spend all our time collecting data and reporting to funders. When we did perform analytics, it would take days and weeks to slice and dice the data. Fast forward to 2017 and we have capacity for quick turnaround of data dashboard reports and other analytics to help inform decision making." ~Jackie Anderson, Division Manager, Snohomish County Human Services.
Example dashboard showing income analysis
"The process often starts with finding something unexpected or enlightening in the data. I bring that information to my supervisor and we start reaching out to planning committees and providers to discuss the data. It's really exciting to be in a community that we can bring up these issues and people are ready to engage." ~Jesse Jorstad, Lead Data & Program Analyst for HMIS, Snohomish County Human Services.
The Snohomish County HMIS Team Discussing Options with a Data Visualization Consultant

Data and Veterans Initiative

Increasing capacity for rapid re-housing wasn’t the only decision the dashboards helped the CoC make. During the PIT count in 2013, 99 veterans were identified as lacking stable housing, sparking a new initiative called House the 99. Community partners, including county government, workforce development, local nonprofits, the state VA, elected representatives, and Naval Station Everett, came together to form the Veterans Homelessness Committee. Over three years, the committee housed over 446 veteran households, but the veteran PIT count remained consistent at about 70. Clearly a piece of the puzzle was missing.

Snohomish County’s coordinated entry system offered a new opportunity for the HMIS team to create a dedicated veterans’ dashboard that showed, for the first time, the net impact of inflow and outflow of veterans, exposing an underlying problem of veterans becoming homeless at the same rate they were being housed. As a result, the CoC enhanced the monthly by-name list and began assessing new potential policies, in partnership with veteran programs and community stakeholders, to stem the flow of veteran entry into homelessness and address the lack of housing for veterans across the County.

Example charts from the HMIS showing the number of homeless veterans currently in the coordinated entry records in the HMIS and how many veterans the HMIS shows are housed
Example of charts from the HMIS showing veteran inflow/outflow rates in the system
Presentation at the Affordable Housing Breakfast Forum

Data Quality Improvement

The dashboards and their stories have greatly impacted data quality.

Snohomish County CoC already had a comprehensive data quality management plan - improving data quality did not require alterations or stricter enforcement to its existing data quality policies. They found that frequent education and exposure to how a provider’s data affects metrics, such as system performance measures, were enough of a catalyst to see improvement in data quality from participating providers. The HMIS team created two kinds of dashboards to help with data quality

  1. Advanced back-end data auditing dashboards
  2. Ad-hoc data quality tables to address specific or provider requested issues

These visualizations allow for timely and streamlined assessment of data quality issues. As questions arise when a provider is reviewing reports, the necessary data fields can be pulled within minutes or hours into an ad-hoc report that provides an explanation and actions that the provider can use to fix the data. This responsiveness has helped providers trust HMIS to provide them with answers to their questions and trust that the data analysis is going to be accurate. The presence of data stories created by the dashboards and easy access to data quality tools help providers see that their data is accurately representing their projects, due to their own data quality efforts.

"If stakeholders don't feel like the data align with their understanding... you can't move forward." ~Jesse Jorstad, Lead Data & Program Analyst for HMIS, Snohomish County Human Services
Example data quality dashboard showing data completeness for universal data elements
Example data quality dashboard for rapid re-housing
Brainstorming Notes

From Someone Who’s Been There

The dashboards have ultimately helped free up HMIS resources.

Staff spend less time writing queries and reports and more time answering questions and providing guidance on how to use and understand the data. Due to quick turnaround of data requests, the HMIS team has found that providers are much more likely to rely on and translate the data into practice improvements. Here are some lessons that helped the Snohomish County HSD team be most effective:

  • Leverage a replicable standard data export from the HMIS such as the APR/ESG & DQ Export or the Standard CSV export to import into a data visualization software.
  • Clearly define what the HMIS data elements are and how they can be used; identify essential data elements that will be required to build the dashboards and answer questions.
  • Write out standard formulas that are commonly used which can be copied and pasted into the output file, such as those that combine response fields. Grouping response fields help visualize the data into how people think about them. For instance, “Prior Residence” responses can be grouped into categories such as Literally Homeless, Institutional Setting, Permanent Housing, etc.
  • Identify the questions to be answered through discussions with partners before creating dashboards. Let those questions guide the design and queries of the dashboards.
  • Invest in software that can house stored procedures or calculated fields which automatically refresh with an updated source so dashboards can be “plug and play.”
  • Make the data accessible and consistent – create familiarity with consistent representation of data.
  • Build sustainable relationships with community partners.
  • Insert the HMIS team into the conversation whenever possible; make the team the go-to place for understanding data when discussing policies and strategies.
  • Build trust through answering questions and walking people through the data.

The Snohomish County HMIS team warns against being too focused on the initial vision of the end-product when starting the process. Dashboards should include continuous evaluation, improvement, and dynamic reporting to ensure questions continue to get answered. Creating static dashboards limits the questions that can be answered.

"The biggest gift of this entire process is that we have a way of processing data and meeting data requests that is wicked fast." ~Jesse Jorstad, Lead Data & Program Analyst for HMIS, Snohomish County Human Services
Jesse Jorstad, Lead Data & Program Analyst for HMIS, Snohomish County Human Services presenting findings to the Partnership to End Homelessness on the relationship of System Performance Measures

Next Steps as a Data Culture

Snohomish County HSD is looking forward to continuous evaluation – and data are the foundation for answering their questions. The dashboards are creating opportunities for the CoC to more effectively evaluate performance and develop appropriate system and project goals. From designing project type-specific measures to reconciling PIT count data with coordinated entry data, Snohomish County is leveraging dashboards to demonstrate the housing system is more efficient and effective than ever before while continually identifying areas for improvement. They are digging into the data.

And the data are bringing people together.

Jesse describes “a palpable moment of understanding that, in the struggle to end homelessness, we were losing ground.” He adds, “It took a long time of looking at the data as a whole to understand that homelessness presents itself differently for various populations. The same goes for providers. They may have a lot of knowledge in a specific part of homelessness, but when trying to make system changes, we need to look at it as a whole. Now everyone is working towards the same goals instead of solely meeting the requirements of their specific part of the system.”

Example infographic showing progress using inflow/outflow of households in the system

Snohomish County HSD and their partners have used the dashboards and the expertise of the HMIS team to create a culture of strategic planning that incorporates and is educated by data. They plan to continue using data visualizations to establish a sustainable framework for continuous reporting, evaluation, and translation of data into system improvements. The CoC developed a strategic plan to prevent and end family homelessness that was informed by data, creating a performance culture with community-wide recognition of the value of data-driven decision-making. Snohomish County HSD recognizes the data visualizations as great resources, but they acknowledge that they may only be useful in communities that are willing to be open to data exploration, have community partner buy-in, and if they are built with tools that are sustainable and dynamic. Snohomish County has shown that with these key pieces in place, effective system transformation can begin.

Example of the revised dashboard depicting number of all households, length of time in project, exits to permanent housing, permanent housing retention, days to move-in, Income progression, and returns to homelessness
"The key to building understanding does not lie in the complexity of the analysis, but in the consistency with which it is executed and the certainty of what is being measured and what is not. So, go forth and measure something! Anything! And then share it with people. When they ask more questions – you know you have made an impact." ~Jesse Jorstad, Lead Data & Program Analyst for HMIS, Snohomish County Human Services
Jesse Jorstad, Lead Data & Program Analyst for HMIS, Snohomish County Human Services, showing a line graphic depicting the increased value of presentation skills over time