Unsupervised NLP for Industrial Support Ticket Analysis: SCM Group Case Study in Text Clustering and Topic Modeling
(2025) DABN01 20251Department of Economics
Department of Statistics
- Abstract
- This thesis was developed in partnership with SCM Group, a leading Italian manufacturer of material-cutting machinery operating an extensive international network of production sites and clients. Because of the magnitude of its operations, SCM’s technical support teams manage a large volume of support tickets-records submitted by clients on a daily bases: structured reports documenting machine failures, requests for intervention, or technical inquiries.
The company’s historical ticket data is a valuable resource. This project uses unsupervised machine learning and Natural Language Processing (NLP) to extract actionable insights from the tickets. These insights include: Trends in maintenance demands to guide proactive quality improvements;... (More) - This thesis was developed in partnership with SCM Group, a leading Italian manufacturer of material-cutting machinery operating an extensive international network of production sites and clients. Because of the magnitude of its operations, SCM’s technical support teams manage a large volume of support tickets-records submitted by clients on a daily bases: structured reports documenting machine failures, requests for intervention, or technical inquiries.
The company’s historical ticket data is a valuable resource. This project uses unsupervised machine learning and Natural Language Processing (NLP) to extract actionable insights from the tickets. These insights include: Trends in maintenance demands to guide proactive quality improvements; Recurring issue patterns (e.g., frequent faults linked to specific machine models or components); and Automated topic classification using clustering techniques (e.g., BERTopic for interpretable clusters, K-means on TF-IDF vectors) to streamline ticket prioritization.
Key challenges include sparse text in critical fields and multilingual content; nonetheless, the suggested pipeline- text preprocessing, feature engineering, and cluster-trend analysis- converts raw ticket data into a structured knowledge base. The findings enables SCM Group’s decision makers to better understand common hurdles, detect trendsacross components and machine models, and ultimately enhance the company’s service and product quality. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/9192523
- author
- Naletto, Elide LU
- supervisor
- organization
- course
- DABN01 20251
- year
- 2025
- type
- H1 - Master's Degree (One Year)
- subject
- keywords
- Technical Support Tickets, Natural Language Processing, Topic Modeling, Clustering, Knowledge Extraction
- language
- English
- id
- 9192523
- date added to LUP
- 2025-09-12 09:04:57
- date last changed
- 2025-09-12 09:04:57
@misc{9192523, abstract = {{This thesis was developed in partnership with SCM Group, a leading Italian manufacturer of material-cutting machinery operating an extensive international network of production sites and clients. Because of the magnitude of its operations, SCM’s technical support teams manage a large volume of support tickets-records submitted by clients on a daily bases: structured reports documenting machine failures, requests for intervention, or technical inquiries. The company’s historical ticket data is a valuable resource. This project uses unsupervised machine learning and Natural Language Processing (NLP) to extract actionable insights from the tickets. These insights include: Trends in maintenance demands to guide proactive quality improvements; Recurring issue patterns (e.g., frequent faults linked to specific machine models or components); and Automated topic classification using clustering techniques (e.g., BERTopic for interpretable clusters, K-means on TF-IDF vectors) to streamline ticket prioritization. Key challenges include sparse text in critical fields and multilingual content; nonetheless, the suggested pipeline- text preprocessing, feature engineering, and cluster-trend analysis- converts raw ticket data into a structured knowledge base. The findings enables SCM Group’s decision makers to better understand common hurdles, detect trendsacross components and machine models, and ultimately enhance the company’s service and product quality.}}, author = {{Naletto, Elide}}, language = {{eng}}, note = {{Student Paper}}, title = {{Unsupervised NLP for Industrial Support Ticket Analysis: SCM Group Case Study in Text Clustering and Topic Modeling}}, year = {{2025}}, }