Emotional Analysis of Fashion Trends Using Social Media and AI: Sentiment Analysis on Twitter for Fashion Trend Forecasting

TL;DR

This work demonstrates how Twitter-based sentiment analysis, coupled with time-series decomposition and causal inference, can forecast fashion trends. By integrating a seven-theme taxonomy, improved sentiment normalization, ARIMA/SARIMA forecasting, Granger causality, and cross-platform/brand analyses, the study delivers statistically validated insights into trend dynamics and drivers. Key findings include statistically significant rises in accessories and streetwear, a central role for sustainability in the causal network, and pronounced platform and brand differences, all supported by a balanced predictive model with strong performance metrics. The approach provides actionable early indicators for brands and retailers to anticipate emerging fashion trajectories with quantified confidence.

Abstract

This study explores the intersection of fashion trends and social media sentiment through computational analysis of Twitter data using the T4SA (Twitter for Sentiment Analysis) dataset. By applying natural language processing and machine learning techniques, we examine how sentiment patterns in fashion-related social media conversations can serve as predictors for emerging fashion trends. Our analysis involves the identification and categorization of fashion-related content, sentiment classification with improved normalization techniques, time series decomposition, statistically validated causal relationship modeling, cross-platform sentiment comparison, and brand-specific sentiment analysis. Results indicate correlations between sentiment patterns and fashion theme popularity, with accessories and streetwear themes showing statistically significant rising trends. The Granger causality analysis establishes sustainability and streetwear as primary trend drivers, showing bidirectional relationships with several other themes. The findings demonstrate that social media sentiment analysis can serve as an effective early indicator of fashion trend trajectories when proper statistical validation is applied. Our improved predictive model achieved 78.35% balanced accuracy in sentiment classification, establishing a reliable foundation for trend prediction across positive, neutral, and negative sentiment categories.

Figures (13)

• Figure 1: Comparison of original (left) and improved (right) sentiment score distributions, showing the effect of enhanced normalization in reducing extreme values and providing more balanced sentiment categories.
• Figure 2: Time series decomposition for sustainability fashion theme showing trend, seasonal, and residual components.
• Figure 3: Time series decomposition for luxury fashion theme showing trend, seasonal, and residual components.
• Figure 4: Time series decomposition for vintage fashion theme showing trend, seasonal, and residual components.
• Figure 5: Fashion theme trends with statistical significance indicators, showing confidence levels and differentiating between statistically significant and non-significant trends.