Automated Political Information Assistant

An AI-powered tool to automate the processing of political information from the Dutch House of Representatives.

This project aimed to develop an automated assistant for processing political information, addressing inefficiencies in handling debates and publications from the Dutch House of Representatives. Previously, manual tasks such as monitoring the agenda, transcribing debates, and linking documents were time-consuming and inconsistent.

The solution we built automates processes to save time, enhance reliability, and improve information accessibility.

The Solution

The tool we developed integrates daily monitoring of relevant websites, such as tweedekamer.nl and rijksoverheid.nl, to gather and process key information. It enabled users to access summaries, and a grounded interactive chatbot for real-time Q&A, providing both speed and accuracy. A central knowledge base was created to store all data, making it searchable by topics, dates, or keywords for comprehensive data management.

The final deliverables included:

  • A functional AI tool capable of gathering information from different sources.
  • Generation of summaries at various levels of detail.
  • An up-to-date, organized knowledge base.

My Role and Technologies Used

As one of two developers on this project, I was co-responsible for the end-to-end development of the solution. This involved working with a modern tech stack to deliver a robust and scalable application.

Backend Development

The backend was built using .NET 9 and ASP.NET Core Web API, providing a high-performance and scalable foundation. Key components included:

  • Azure AI Services: We leveraged a suite of Azure services to power the AI capabilities:
    • Azure OpenAI for generating summaries.
    • Azure Cognitive Search for the intelligent search functionality.
    • Azure Cosmos DB for storing documents.
    • Azure Cognitive Speech Services for transcribing debates.
  • Web Scraping and Automation:
    • Playwright for robust web scraping.
    • Quartz.NET to schedule daily data gathering tasks.
  • Data Processing:
    • PdfPig to extract text from PDF documents.
    • YoutubeDLSharp to process video content from debates.

Frontend Development

The frontend was developed using Next.js, React, and TypeScript, creating a fast and modern user experience.

  • The UI was styled with Tailwind CSS.
  • SignalR was also used on the frontend to receive real-time updates from the backend.

What I Learned

This project was a great opportunity to deepen my expertise in building AI-powered applications using the latest Microsoft technologies. I gained valuable experience in:

  • System Architecture: Designing and implementing a complex, distributed system with multiple moving parts.
  • Azure AI and Cognitive Services: Integrating various Azure services to build a powerful and intelligent application.
  • Full-Stack Development: Working on both the .NET backend and the Next.js/React frontend, giving me a holistic understanding of the entire application.
  • Automation and Data Processing: Implementing robust automation and data processing pipelines to handle large volumes of data from various sources.

How My Skills Can Help Your Business

The skills and experience I gained from this project are highly transferable and can be applied to a wide range of business problems. If you are looking to:

  • Automate manual processes and improve efficiency.
  • Build intelligent applications that leverage the power of AI.
  • Develop scalable and robust web applications with a modern tech stack.
  • Integrate with various third-party services and APIs.

… then I have the expertise to help you achieve your goals. Please feel free to contact me to discuss your project.