Simplifying Dashboard Development Through a Domain-Specific Language: A Web-Based Approach

Abstract

This dissertation presents the design and development of a web-based platform aimed at simpli- fying the creation of data dashboards through a domain-specific language (DSL). The proposed solution targets users with limited technical background or programming experience, enabling rapid and flexible configuration via human-readable YAML files without requiring compilation or build steps. At the core of the system lies a modular and extensible DSL for defining dashboard compo- nents-such as charts, data sources, filters, and aggregation operations. An interactive web editor was developed to allow users to modify YAML configurations and preview the dashboard in real time, facilitating immediate visual feedback and reducing iteration cycles. The platform supports a wide range of chart types (e.g., bar, line, pie, radar, stacked bar), data transformations (e.g., sum, average, count), and integration with both CSV files and PostgreSQL databases. The backend, built with Flask (Python), handles configuration parsing, version control, and file management. The frontend, developed in React, leverages Chart.js for rendering and AceEditor for YAML editing. It includes features such as user authentication, soft deletion, historic version tracking, syntax validation, and a continuous WYSIWYG preview mechanism. Usability testing was conducted with participants of varying technical proficiency, combining scenario-based observation with the System Usability Scale (SUS). Results indicate that the plat- form effectively supports users with limited technical skills in creating and iterating dashboards, validating the feasibility of DSL-driven dashboard generation in practical settings. This work contributes to the field by proposing a dynamic, extensible, and user-centered ar- chitecture, with the potential to be adapted across various domains seeking to democratize access to data visualization. ii