I’m building an automated, end-to-end pipeline that pulls results from a fast-changing election website every few minutes, cleans and enriches the feed with AI, then pushes out clear charts and graphs that highlight vote counts per party, regional voting trends, overall voter turnout and a concise statistical summary of the outcomes. There is an existing Google Data Studio (Looker) template available. The scope breaks down into three tightly-linked steps: • Data capture – a headless, resilient scraper (Selenium, Playwright or a similar tool) must track roughly 3 500 individual race entries across 17 political parties in six regions, coping smoothly with AJAX calls, pagination and any CAPTCHA or session refreshes. • AI-powered processing – once ingested, the stream needs real-time de-duplication, error checking, normalisation of party names/regions and basic predictive analysis. Leveraging APIs such as OpenAI, Claude or Vertex for cleaning and pandas / scikit-learn for statistical work is fine; I’m happy to fund premium tiers where that improves reliability and speed. • Visual output – the cleaned dataset should feed straight into a powerful dashboard ( Power BI, Tableau, Florish, Looker Studio, Streamlit, Plotly Dash, Superset or similar) that automatically refreshes and renders laser-sharp bar charts, line charts and comparative graphs. Quick download options for CSV/Excel will be required. Acceptance criteria 1. A single command or deploy-to-cloud button stands up the full workflow. 2. Scraper demonstrates <1 min lag behind the source during live test. 3. Accuracy benchmark: ±0.2 % versus manual spot checks on 50 random records. 4. Final dashboard loads in under five seconds on a standard laptop and updates without manual refresh. If you’ve delivered mission-critical scraping or live analytics before and can show sample dashboards, I’m ready to start immediately. Project set up and demonstration cannot exceed 2 days