Apify Social Listening Pipeline Build

Замовник: AI | Опубліковано: 11.11.2025
Бюджет: 3000 $

I’m setting up a full-stack, Apify-driven social-listening workflow focused on Caribbean politics and need a specialist who can take it from data collection to insight-ready dashboards. Scope The pipeline must cover Facebook and Instagram. On Facebook pages we need posts, comments, likes, shares, user profiles, comment timestamps, anonymised user IDs and full engagement counts. Instagram should follow the same logic for public posts. Everything has to run automatically each day inside Apify, with retry logic and clear logging. Data flow & analysis Raw JSON or CSV should land in an Apify dataset, flow into a durable database (BigQuery, Postgres or similar) and trigger analysis. The analysis layer must surface sentiment (positive / negative / neutral), engagement trends, keyword frequency, issue clustering and narrative mapping, influencer and account impact tracking, plus early-signal virality detection. Feel free to draw on open-source NLP libraries—Hugging Face Transformers, spaCy, Vader, or your preferred stack—so long as the models handle Caribbean English and local dialects well. Dashboards Interactive visuals (Tableau, Power BI, Data Studio or a lightweight web app) should let us filter by keyword, sentiment, date range, account and geography, with an export-to-CSV option for every chart. Deliverables • Apify actor(s) for Facebook & Instagram, authenticated and rate-limited, scheduled daily • ETL scripts / workflows pushing data into the analytics store • Analysis notebook or service performing the metrics listed above • Live dashboard linked to the analytics store • Step-by-step deployment & handover guide, including how to add new keywords, pages or actors Acceptance criteria 1. A 24-hour unattended run populates the database with fresh posts and metrics. 2. Sentiment precision ≥ 80 % on a 200-item validation set supplied by me. 3. All dashboard filters load in <3 seconds on a standard broadband connection. 4. Clear documentation lets a non-dev teammate rerun or extend the pipeline within 30 minutes. If this end-to-end build is in your wheelhouse, let’s start planning the actor structure and tech stack right away.