Social Sentiment Analysis Application Build

I already have a working machine-learning model that detects sentiment on social media for Bank Mandiri, and I now need it turned into a full-fledged application. The goal is simple: pull posts from X (Twitter), Instagram, Threads, and TikTok, classify every entry as positive, negative, or neutral, and display clear trends and summaries on an interactive dashboard so non-technical stakeholders can see the overall mood at a glance.

You will start from the existing notebooks and training data I provide, connect the relevant platform APIs, wrap the model in production-ready code, and create a responsive dashboard. I am comfortable with Python, so a stack that combines Python (FastAPI or Flask) on the back end with Streamlit, Dash, or a lightweight React front end would fit well, but I am open to alternatives if they speed development without compromising maintainability.

Deliverables
• End-to-end data pipeline that ingests posts from the four platforms in near real time.
• Sentiment classifier that outputs positive, negative, or neutral labels with logged confidence scores.
• Web-based dashboard showing daily and weekly trends, top keywords, and the current sentiment breakdown.
• Setup scripts and concise documentation so the app can be deployed on our internal servers or a cloud VM in one step.

Acceptance criteria
• Accurate classification demonstrated on a held-out validation set I supply.
• Dashboard loads in under five seconds with one month of data.
• All services containerised (Docker) and runnable via a single compose file.

If you have built similar analytics tools or have experience with the X, Instagram, Threads, or TikTok APIs, I would love to see your work. Let’s turn this prototype into a robust application that brings sentiment insights right onto our screens.