Oblisk is a long-term solo project. The idea emerged gradually and matured through several design and specification phases. Development has started and is currently ongoing. The project is designed as a full SaaS platform, with a strong emphasis on UX, security, and collaborative features.

The goal is to design a professional-grade product, suitable for both technical and non-technical users, with a high level of reliability and care in both the UX and the backend architecture.

There are no external stakeholders involved, but the project was shaped by multiple research rounds, persona definition, and prototyping sessions. My primary objectives were to sharpen my skills in product architecture, design systems, collaborative UX patterns, and secure data handling. I also wanted to challenge myself to define and scope a product from scratch, including both frontend and backend requirements.

Problem and Context

Accessing, exploring, and sharing data remains complex, especially for small teams, freelancers, or non-technical users. Existing tools are often built for large organizations, require licenses, and introduce friction for collaboration. In many cases, they force data to be uploaded or stored externally, which creates trust and security concerns.

Oblisk was designed to address this by focusing on three core needs:

→ Simplified access to data for all user levels.
→ Seamless collaboration through shared canvases and dashboards.
→ Strong control over data privacy, especially for sensitive or internal data.

Most competing tools fall short in one or more of these areas. Some are too technical, others too rigid or closed. Oblisk aims to fill the gap by making data truly usable and shareable, without compromising security or usability.

I wrote an article on the importance of data visualization here

Process and Choices

Architecture and Stack

The backend relies on Supabase, used for authentication, PostgreSQL database management, and file storage. For custom logic and future extensibility, a separate NestJS backend is planned, written in TypeScript and using REST APIs where needed. Prisma may be used in the standalone backend version, but is not required in the Supabase-managed stack.

The frontend is developed in React (TypeScript), using Next.js, Tailwind CSS, and Zustand for state management. Animations are handled via GSAP. Data visualization relies on D3.js for 2D and R3F for 3D, with Vizx considered for chart primitives. The stack is organized as a monorepo (Turborepo).

Authentication is handled via Supabase Auth (JWT), and query caching is done via Tanstack Query. Monitoring is planned through Sentry. Data querying uses live connections rather than data duplication, following practices used in tools like Metabase or Retool.

Data Model and Handling

Oblisk supports file imports (CSV, JSON, Google Sheets) and live queries to databases. Data is not stored permanently unless explicitly imported as a static file. Credentials for live connections are encrypted at rest and in transit. A short-term in-memory cache is available for performance but never persisted.

The system stores metadata about datasets, user permissions, canvas state, and dashboard structure. These are all centralized in a Postgres database and exposed through versioned APIs.

UX Principles

The design philosophy is to create a smooth, modern, and collaborative data experience. The interface avoids rigid dashboards and aims to be fluid, responsive, and cleanly designed. A dedicated design system ensures consistent use of spacing, typography, motion, and components. Visualizations follow strict rules for legibility and layout, and are designed to be shareable without modification.

Features are introduced progressively, and advanced options are hidden until required. The platform is designed to be used by non-technical users without any onboarding. Visual customization is encouraged, and graphs can be exported in multiple formats (PNG, SVG, GLB, PDF…).

Collaboration is built-in: users can annotate graphs, work in real-time on shared canvases, and communicate via contextual comments. An AI-powered assistant is available but used discreetly, to avoid overwhelming the interface or breaking trust.

Personas

Oblisk was designed with specific personas in mind, which shaped many of the feature decisions. Below are a few representative profiles.

Karim, 29 – Data Engineer in a mid-size logistics company

Karim manages the company’s data stack. He needs light, efficient tools that he can deploy quickly without security risks. He doesn’t want data leaving the internal network and avoids vendor lock-in.

Needs:

→ Live database connection, without external data storage
→ Quick views to debug or present data
→ Easy integration in internal infrastructure
→ Clean UI for non-technical teammates

Léa, 23 – Freelance data consultant

Léa works with small businesses and associations. She creates one-off reports and needs a simple, visual tool to work fast.

Needs:

→ Easy CSV or GSheet import
→ Clean infographics for client reports
→ Google login
→ Free tier with enough features for small projects

Ema, 27 – Product Owner in a tech startup

Ema bridges tech, design, and business. She needs to understand insights from the data team, work with them to define product priorities, and share results internally.

Needs:

→ A shared space to explore insights collaboratively
→ Annotations, comments, and shared dashboards
→ High-level summaries for C-level meetings
→ Reuse of visuals in internal communications

Camille, 32 – Independent consultant in organizational diagnostics

Camille helps companies analyze internal processes and team dynamics through surveys and structured interviews. She often receives raw data from HR or operations teams but lacks the tools to turn them into clear and reusable visuals. The support she gets from technical staff is limited, and she frequently has to reformat everything manually before presenting it to clients.

Needs:

→ Clear visualization of survey results
→ Adapt visuals for operational or executive-level audiences
→ Add comments and contextual notes
→ Export clean visuals for reports, slides, or workshops

Adrien, 31 – Ops manager in an e-commerce brand

Adrien handles logistics and needs access to operational KPIs. Current internal tools are too rigid and require developer help.

Needs:

→ Build and update dashboards without tech support
→ Explore data freely
→ Share results with team easily

Other secondary personas include designers, educators, artisans, and small business owners. All share the same need for an intuitive and secure way to make sense of their data.

Results

Oblisk is still in the early stages of development. The product has been extensively scoped, with detailed specifications, UX principles, and architecture plans. The component library and visual direction are also being documented separately.

The platform is not yet public, and no functional version has been released.

Current Perspective

I plan to continue development in parallel with other work, and eventually release a working public version. Oblisk is a long-term product foundation.