I’m building a real-time, AI-driven dashboard that will automatically score and segment B2C leads for marketing campaigns. The core objective is to surface the highest-value prospects by analysing three metrics side by side—lead source, conversion rate and customer lifetime value—then presenting clear, actionable insights in an interactive interface. Here’s how I picture the flow. Raw data will stream in from typical marketing touchpoints (landing pages, paid ads, email, social, CRM exports). A lightweight ETL layer will clean and unify these feeds, feed them into a predictive model that ranks each incoming lead, and push the processed results to a front-end dashboard. I’m open to frameworks like Python + FastAPI for the back end and React, Vue or similar for the UI, so long as the final product feels quick, modern and easy for non-technical marketers to navigate. Key deliverables • A reproducible data pipeline that ingests and normalises lead files or API feeds • A machine-learning model (e.g., scikit-learn, TensorFlow) that outputs a lead quality score with transparent feature weighting • A web-based dashboard that filters by lead source, visualises conversion rates over time and projects customer lifetime value in currency terms • Documentation covering environment setup, model retraining steps and a short user guide for my marketing team Acceptance criteria 1. Dashboard loads in under three seconds with a test data set of 50k records. 2. Model shows ROC-AUC ≥ 0.80 on a held-out validation set. 3. All three metrics (source, conversion, CLV) update live when new data is pushed via API. 4. Codebase is version-controlled (Git) and ready for deployment to a cloud host (AWS, GCP or Azure). If this aligns with your expertise, I’d love to see a concise outline of your proposed tech stack and timeline so we can move straight into building and iterating.