This past summer The University of Washington went live with Workday, a Software-as-a-Service human capital management solution. It replaced a decades-old HR/payroll (HR/P) system and represented the largest administrative transformation in the school’s history.
Human resources personnel are affected by the change, of course, but so too is practically everyone else on campus. That’s to the tune of some 30,000 people, including employees and managers who input and approve time off using the system, or others who access its data through the Enterprise Data Warehouse (EDW) for analysis.
LIVE ONLINE TRAINING: ONTOLOGY ENGINEERING
Get an overview of ontology engineering skills needed for knowledge graph applications during this one-day live course – Nov. 30, 2022.
“It’s a huge change,” says Pieter Visser, solutions architect at the University, both in the fact that Workday is Cloud-based and that it operates differently than the old system from a data perspective. How to make that huge change with greater ease? The idea was to create Knowledge Navigator (KN) to facilitate the system migration: Based on the Neo4j Graph Database, Knowledge Navigator is an enterprise metadata repository tool for administrative data and change management.
“We use that metadata repository to help people understand the changes, especially on the EDW side – what is the impact to them and how they could prepare for those changes,” Visser says. For instance, Knowledge Navigator can interpret table and column changes that took place as part of the systems migration, so even though users may not be able to get the exact same report they were used to in the old HR/P solution, they should be able to find similar data in another report or in Workday itself.
Workday bills itself as a disruptive technology, says Matt Portwood, the product manager for Knowledge Navigator – one that includes changing culture. The language and processes people used in the previous HR/P system don’t exist anymore: Job classifications, for instance, are now called job profiles, while person is replaced by employee, which is a worker type. “How to adapt to this new business culture is a core part of what the Knowledge Navigator system particularly was aimed at,” Portwood says. “You have definitions and examples and deeper information to follow those new concepts.”
Metadata Management Matures
Knowledge Navigator is the outgrowth of work the University has been doing for several years in developing metadata Management capabilities. The comprehensive management of metadata, it believes, is vital to enabling an organization to oversee changes while delivering trusted, secure data in a complex data integration environment. Solid Metadata Management tools, the University confirms, play a central role in holistic system management, including system migration.
“This is an add-on to our Metadata Repository,” says Visser. “We had all this metadata for the EDW in our metadata application and then we realized we can use that same data to show the migration within the app.”
As Portwood states, the motto on their team is that unless producers and consumers of information agree on meanings, the information itself is meaningless. “But what Neo4j plus the Knowledge Navigator user interface gives us is the ability to navigate across business terms that live in the glossary and technical metadata,” he says. So, as an example, it’s possible to find agreed-upon definitions of terms, such as STEM, and actually navigate to the data in the EDW that will accurately answer a question like: ‘How many students that graduated with STEM degrees last year were women?’
Neo4j brings to the table its prowess at graphing and visualizing data relationships to further inform Knowledge Navigator results. For instance, Visser notes, it is textually and visually presented that the Person data table that was included in a former report now extends across three different tables in Workday, and might also be found in a different report. Adding relationships and how they go together is “the strength and power of how we give context,” he says.
“And when we talk about metadata, context is key.” The value-add of Neo4j, says Portwood “is its optimization for graphing and visualizing the relationships. You navigate from the business concepts down to technical data where they can explore answers to questions.”
Knowledge Navigator’s Next Job
By highlighting linkages between the old and new systems in a web-based, interactive platform, Knowledge Navigator keeps users informed and engaged throughout the enterprise system migration by providing self-service access to conceptual and technical descriptions, definitions, lineage, interactive relationship maps, and impact analysis information, the University notes.
Employees had two years before Workday went live to prepare for it with the help of Knowledge Navigator. Shortly, they’ll be able to put Knowledge Navigator to work preparing for the replacement of another system, the University’s finance software.
“Knowledge Navigator is key for that as well because we will use the same functionality but also will be able to do impact analysis even before that to identify objects that will be changed, and that way we can identify new users that will be impacted as well,” Visser says.
One of the advantages of having the HR/P experience is that a few key concepts from the HR glossary will be replicated in the finance glossary, too, such as those related to paying employees and budgets. “But more than that there’s going to be the framework to work with external partners to identify terms and get agreements,” Portwood says. One Neo4j-related win has had to do with learning how to work through the process of getting agreement among parties, which can be incredibly hard.
The University has started tracking its search histories (especially when results aren’t found) to see whether Knowledge Navigator can serve as something of an institutional Google search engine, too. People are searching business jargon and acronyms, not just standard terms, and the idea is that Knowledge Navigator can be the place where they start.
It’s not necessarily about documenting everything in Knowledge Navigator but creating links to things, such as internal Wiki pages or other system assets, via Knowledge Navigator, Visser says.
“When you can do that based on metadata it’s very powerful – you might query for this table but we might have a Tableau visualization about that too,” he says. “The more information we can add about people and things in Knowledge Navigator, the stronger it really is.”
Taking the Custom Route
Metadata makes the most sense to document in a graph solution because of its ability to let you visually see the relationships between business concepts and data and how data gets reported in reports. It lets you follow the story, as Portwood frames it.
While the University could have just bought an off-the-shelf metadata tool, Visser says that the problem with most of those solutions is that they are created for metadata managers, not end users. With Knowledge Navigator, “when the end user goes in they immediately understand the context and how to use it. Also, key is the amount of customization and integration we can do, including integrating data points into the metadata,” such as who the end user is in the organization for a richer experience. That’s not something you can easily find in an out-of-the box solution.
That said, there’s nothing to stop the more technical users that want to go deep from doing so. “The graph model is interactive and they can click all day long to explore as much as they want, but other users can search, find the definition and just back out,” says Portwood. The solution, he explains, fits the University’s broad user base.
Neo4j, meanwhile, continues to enhance its graph database. It’s Native Graph Platform added analytics, data import and transformation, visualization, and discovery atop the graph database. It also released the preview version of Cypher for Apache Spark (CAPS) language toolkit, building on its unveiling of openCypher a few years ago – an effort to push the graph industry forward by tapping into the open source community and making Cypher’s evolution an open exercise.
Image used under license from Shutterstock.com