Yesterday, Northwoods announced availability of Traverse, a text analytics solution for child welfare social workers. The solution is designed to allow case workers to focus more on a child’s needs rather than needlessly reviewing case documents. Traverse sifts through a case’s thousands of documents to give case workers quick insights into the people and key topics related to the child. This allows the case worker to focus on the child’s future safety rather than researching the case.
About Northwoods Social Services Solutions
Northwoods can be described in one statement by putting Lee and my favorite homespun vernacular together. The company started with “two guys and a computer” who wanted to do something that “wasn’t putting toilet seats on the internet.” Gary Heinze and Rich Diers wanted to make a difference and decided to focus their skills on human services agencies. They created a centralized document repository for case workers which runs on top of Hyland OnBase, Laserfiche Rio, or IBM’s FileNet. The solution moved agencies from paper into the digital world and is currently used in over a hundred human services agencies around the country. Four years ago, Northwoods brought mobility to case workers and showed it to one of their clients, Rich Bowlen. Rich would join the team and together the three created Traverse, which was announced yesterday.
Putting Text Analytics to Case Work
Here’s how Traverse works. Each case file contains on average 4,000 to 5,000 pages that would typically take several days for a case worker to review entirely. The case worker would need to identify every individual, note every event, and highlight interesting topics. Not only was this process slow but it could be easy to miss a subtle note within a few pages. This becomes even more challenging when you’re the case worker on call coming on the scene in the middle of the night on a case that’s not yours. Traverse gives the case worker instant insights into the individual case files.
The system uses text analytics from IBM Watson to derive information from the documents in the case file. The analytics engine performs entity and topic extraction, defines connections, and creates timelines. But text analytics tools don’t work alone; it needs expertise to configure them. Northwoods, in the role of data scientists, brings domain expertise in child welfare and the understanding of what text analytics can offer. This expertise is used to filter and further refine the results returned from the text analytics tool and develop the data visualizations to present them to the case worker.
For example, the system reviews the documents for individuals involved in the case. It identifies the relationship between each individual and the child, as well as the strengths or safety concerns with each person. The system quickly allows the case worker to review a list of individuals and the number of mentions in that case. By clicking on that individual, the system presents a word cloud of all the topics associated with the selected individual. The system also shows both a timeline of events and a list of documents associated to that individual and that specific topic. This reduces case file review from days to hours. In the commercial world we would state this as dollar savings, but in social work this is represented as the ability to support more people in need.
If the time savings was not enough, Traverse also allows the department heads to see the big picture like never before. The results of the text analytics are used to expose topics or events that have been commonly identified across several case files. For example, identification of a possible increase in opiate use within the community after identifying it as a key topic in several cases.
Food for Thought
Northwoods has developed a great example on where to implement text analytics into an existing content solution. The solution shows how even some simple insight can offer powerful benefits and time savings. It’s not difficult to see that similar solutions could be implemented in other areas of state or local government, like other citizen services or law enforcement. Even the corporate world could use solutions such as this for human resources. As more examples like this become visible hopefully we can stop talking about text analytics and start implementing it.
