Automated social media analysis for urban tourism cities: a low-code workflow approach using the #TravelTuesday campaign

This study develops and evaluates a reproducible low-code workflow for automated emotion analysis of multimodal social media content in an urban tourism context. Using the #TravelTuesday campaign as an illustrative case, it examines how Instagram text and TikTok audio content can be collected, transcribed, and analysed to inform destination image monitoring and digital city-branding decisions by destination management organisations (DMOs). The methodology integrates the low-code workflow automation tool n8n with Apify-based web scraping, Google Sheets, automatic speech recognition (ASR), and AI-based emotion classification via OpenRouter using GPT-5.2. The workflow collects hashtag-based posts, comments, and audio content from Instagram and TikTok and classifies emotional expressions into seven categories: joy, sadness, anger, fear, surprise, disgust, and neutral. The outputs are compared with the pretrained DistilRoBERTa model to assess methodological consistency. The results show that the automated low-code workflow can support scalable emotion analysis across text and audio formats. In the #TravelTuesday case, positive emotions predominated and engagement peaked on the day before the official campaign date. These descriptive patterns illustrate how the workflow can identify temporal variations relevant to campaign monitoring in tourism cities. The study’s principal contribution is methodological. It provides a reproducible low-code workflow for operationalising the emotional dimensions of user-generated content in a city-oriented tourism campaign. The workflow can offer practical guidance to DMOs seeking to monitor campaign performance, identify potential reputational concerns, and inform digital city-branding strategies.