Digital transformation demands a highly efficient platform for project work, business intelligence and enterprise insights. Unfortunately, the tool that many of us use to do the actual work - spreadsheets - hasn’t changed significantly since the 80s. Making matters worse, the business data are usually scattered all over various tools.

Legacy tools are less valuable for today’s modern business because the way teams work has dramatically shifted over the last three decades. In just a few short years, millions of workers have started to collaborate by utilizing some of the many different tools available. However, using multiple tools and different apps is problematic.

The goal of a newly created open-source project management tool is to provide a single solution to support current business models. This means providing various views of the same data to allow smooth communication, prevent duplication of effort and allow moving work forward in a seamless manner.

The goal of this thesis is to develop a data filtering library, that:

• works with any supported user data
• can be used both on backend and frontend (i.e. supports usage from Java and JavaScript)
• can be easily maintained and extended
• provides a reasonable performance
• supports complex comparisons between user data fields (i.e. not just $data > constant, but also $data1 > $data2)

One complex problem is the ability to pass in data structures from both supported languages (Java and JavaScript), and to filter data based on their visual representation. The date data type provides a major challenge. For example, a date is stored as ISO string or milliseconds since epoch in the database. While users might have a custom formatting that displays the date as quarters, e.g. 2020 Q2. Such a representation contains string ‘Q2’. The date therefore contains ‘Q2’.

Also performance should be considered. As a baseline, the expectation is to filter 10,000 records in one second on a single core virtual CPU.

To allow multi platform support with the ability to update the source code only once, the code should be written in Kotlin.

The student needs to run a proof of concept of cross compiling the Kotlin code, using language native collections in the source code, and to call native JS libraries. Next, the student needs to develop the needed filtering code, verify its performance and search for possible improvements using a profiler.