Communication between Domain Experts and Data Analysts
It has become increasingly easy to collect huge amounts of data. Making use of this data hasn’t become much easier, though. To make sense of data, you need to combine two things: knowledge about the specific domain the data originates from (e. g. medicine) and methodic knowledge about data analysis.
Connecting these two requires tight communication between domain experts and data analysts. Since the problem domain is often very complex, facilitating this communication with software is the logical next step. This is why Glanos created DataSphere, an application to structure the communication process between domain experts and data analysts.
SCREENSHOT OF A MODULE THAT CONTROLS AND EVALUATES TEXT EXTRACTION, USING WIKIPEDIA AS AN EXAMPLE
Domain Experts, Data Analysts and UX Designers
DataSphere bridges the gap between non-technical domain experts (Glanos’ clients) and highly technical data analysts (the team of Glanos) and lets them effectively collaborate on finding answers in the ocean of data. To make this collaboration work smoothly, Glanos teamed up with the UX experts interfacewerk.
The dedicated tools and design processes of the UX experts ensure that DataSphere makes it easy to collaborate on data analysis. As DataSphere is a tool for every day use, it’s important that the users feel safe handling their data and that the domain experts feel empowered to guide the data analysis.
The UX team of interfacewerk made certain of that by using design thinking methods in the concept phase. They also accompanied the development process with usability tests and interviews to make sure DataSphere is easy and efficient to use.
Controlling the Analysis without understanding technical details
Later on in the collaboration between experts and analysts, the domain experts’ knowledge needs to not only guide but also control the analysis algorithms for making use of in in the daily life. To achieve this, the non-technical domain expert needs to understand enough about the analysis process to fine-tune it.
DataSphere provides a layer on top of the complex algorithms to make it easy for domain experts to take control of the analysis. Only a focus on UX allows the DataSphere developers to hit the sweet spot between oversimplification and overcomplication. This sweet spot turned out to be hard to find, because non-technical users don’t understand complex systems well – they’re not used to that. That’s why DataSphere provides easy to understand summaries first and allows the users to drill down into the details on request. This also increases trust in the system, because figuring out the details is always possible.
Domain experts are guided through the decisions they have to make to provide all necessary background information to the data analysts. This process is improved by designed and tested views that provide the right amount of context for each decision. Empowered by the specific input, the analysts can then successfully transform the supplied information into optimal analysis results.
APPROVAL PROCESS FOR TEXT EXTRACTION RESULTS, ALSO SUPPORTING KEYBOARD SHORTCUTS
Understand and check analysis results
After the analysis is finished and results are there, the domain expert then needs to check them and give feedback on what to improve. UX Design makes this process fast: it allows domain experts to quickly differentiate between correct and incorrect results. The data analysts then get this data in a format that’s perfect for them to continue with.
By shortening this feedback circle, the innovative UI of DataSphere (developed with Angular JS) allows for quick iterations and improvements. It turned out to be especially important that each view has powerful and easy to use filtering tools, so each user can access exactly the data they want to see. This allows the domain experts to have real knowledge and control over the data analysis. It also ensures that they get all the answers to the questions they wanted to ask their data.