Website Sentiment Analytics Design

I need a clean, well-documented workflow that takes my raw text data, runs полноценный анализ тональности, and delivers results that can be displayed directly on a website.

Here is what matters most to me:
• The sentiment model must reliably label positive, negative and neutral tones (more granular scales are welcome if it improves insight).
• The output should arrive in two forms: a machine-readable file (CSV or JSON) and an embeddable visual component such as a lightweight dashboard or set of charts that I can drop into a standard HTML page.
• Please keep the solution self-contained—Python (pandas, scikit-learn / spaCy / NLTK) or another open-source stack is fine as long as I can rerun it locally.

Acceptance criteria
1. Script/notebook that ingests my data and produces sentiment labels.
2. Reusable visualization (e.g., Plotly, Chart.js, D3) styled to match a modern website.
3. Brief README explaining setup, model choices and how to update the visuals with new data.

If you have a preferred NLP library or an idea for a slick front-end integration, let me know in your proposal.