Predictive Intelligence Dashboard

An end-to-end predictive intelligence platform designed to turn complex datasets and real-time external signals into actionable forecasts.

I designed and implemented the dashboard interface in Next.js and React, using TypeScript, Tailwind CSS and shadcn/ui on Base UI to create a structured and scalable design system. Interactive charts, data visualisation and filtering make complex predictive information easier to interpret and act upon.

The platform combines historical datasets with live external API data, allowing forecasting models to incorporate changing real-world conditions rather than relying solely on historical patterns.

Alongside the product design and frontend implementation, I established the application architecture, server infrastructure, deployment environment and data layer, integrating the forecasting service into the wider application.

The frontend is built with Next.js, React and TypeScript, with Tailwind CSS, shadcn/ui on Base UI and Recharts forming the core interface stack. The forecasting layer uses Python and Nixtla/statsforecast, incorporating models including Holt-Winters, Theta, AutoARIMA, AutoETS and LightGBM.

Managed infrastructure & hosting

The platform is hosted and maintained within my managed server environment, covering deployment, server configuration and ongoing infrastructure management. Productised bespoke versions of the platform can also be deployed into the same managed environment, giving clients a complete build-and-host solution without requiring them to manage their own infrastructure.