Scholarlink

<p>Sustainable supply chain management increasingly depends on understanding what customers actually value, yet most evidence on sustainable supply chain practices (SSCPs) still comes from surveys or expert judgment and lacks real-time, demand-side granularity. This study addresses this gap by mining Twitter to identify which SSCPs customers discuss most and how those practices co-occur across environmental, economic, and social dimensions. We collected 19,540 tweets posted between 2022 and 2024 using the query (#supplychain OR #sustainability) AND practices, then applied rigorous filtering to isolate customer voices. Following manual annotation of 1,141 tweets with strong inter-annotator agreement (Cohen’s κ = 0.82), we employed a hybrid machine learning framework combining TF-IDF feature extraction, SMOTE for class balancing, and supervised models (logistic regression, SVM, random forest) alongside semi-supervised Transductive SVM to classify the remaining corpus at scale. Network analysis with centrality measures revealed the structure of customer discourse. Results show environmental practices dominate online attention (46.15% of discourse), particularly waste management (27.9%) and green manufacturing (22.0%), signaling strong customer demand for eco-friendly initiatives. Economic practices (33.33%) appear prominently, reflecting expectations for efficiency alongside sustainability, whereas social practices (20.51%) receive less frequent mention. Network analysis identifies waste management as the central hub (degree centrality = 0.354, betweenness = 0.187), bridging environmental, economic, and social domains. These practices form tightly interconnected clusters (modularity Q = 0.68), supporting holistic implementation strategies. Managerially, firms can use social media analytics as an early-warning sensor to prioritize high-salience practices, align sustainability investments with market expectations, and integrate cross-dimensional actions to strengthen competitiveness and stakeholder trust.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Customer-Centric Insights into Sustainable Supply Chain Management: A Machine Learning Approach with Twitter Data

  • 期刊论文

  • 发表日期：

    2026年2月21日

查看全文
  • Mehrdad Maghsoudi,
  • Mohammad Rahimi,
  • Sajjad Shokouhyar

摘要

Sustainable supply chain management increasingly depends on understanding what customers actually value, yet most evidence on sustainable supply chain practices (SSCPs) still comes from surveys or expert judgment and lacks real-time, demand-side granularity. This study addresses this gap by mining Twitter to identify which SSCPs customers discuss most and how those practices co-occur across environmental, economic, and social dimensions. We collected 19,540 tweets posted between 2022 and 2024 using the query (#supplychain OR #sustainability) AND practices, then applied rigorous filtering to isolate customer voices. Following manual annotation of 1,141 tweets with strong inter-annotator agreement (Cohen’s κ = 0.82), we employed a hybrid machine learning framework combining TF-IDF feature extraction, SMOTE for class balancing, and supervised models (logistic regression, SVM, random forest) alongside semi-supervised Transductive SVM to classify the remaining corpus at scale. Network analysis with centrality measures revealed the structure of customer discourse. Results show environmental practices dominate online attention (46.15% of discourse), particularly waste management (27.9%) and green manufacturing (22.0%), signaling strong customer demand for eco-friendly initiatives. Economic practices (33.33%) appear prominently, reflecting expectations for efficiency alongside sustainability, whereas social practices (20.51%) receive less frequent mention. Network analysis identifies waste management as the central hub (degree centrality = 0.354, betweenness = 0.187), bridging environmental, economic, and social domains. These practices form tightly interconnected clusters (modularity Q = 0.68), supporting holistic implementation strategies. Managerially, firms can use social media analytics as an early-warning sensor to prioritize high-salience practices, align sustainability investments with market expectations, and integrate cross-dimensional actions to strengthen competitiveness and stakeholder trust.