Objective
Call center agents in telecommunications companies play a crucial role in the customer experience and the company’s reputation. Their ability to solve problems, provide assistance and offer solutions makes the difference between a positive interaction and an unsatisfactory one that affects customer loyalty. This is why the evaluation of agents is fundamental to guarantee a high quality service aligned with the company’s objectives.
In this research work, a methodology has been proposed using an agent grouping model to differentiate according to their performance. A data set consisting of millions of calls in the last year of MasOrange’s operation has been used to extract key performance indicators for each agent in the company. Dimensionality reduction techniques are applied on these data to transform the initial high dimensional space to a three dimensional space while retaining much of the explained variance. Subsequently, different density-based machine learning clustering algorithms and anomaly detection methods are tested, of which DBSCAN is the one with the best results.
The results obtained indicate that there are two groups of agents, those that are in the normal range and those that perform worse than average. From a business point of view, it is important to understand and educate those agents that are in the second group in order to improve the user experience.
BACHELOR’S THESIS BY:
SANTIAGO KIRIL CENKOV STOYANOV
Academic Experience
- Computer Science and Engineering, Universidad Carlos III de Madrid (September 2021 – September 2025)
Work Experience
- Product Analyst – MasOrange (September 2025 – now)
- Machine Learning Researcher – Universidad Carlos III de Madrid in collaboration with Grupo MasOrange (September 2024 — June 2025)
Technical skills
- Programming languages: Python, C/C++, SQL, HTML/CSS, JavaScript.
- Development libraries: Pandas, Numpy, Tensorflow, Keras, Sci-kit Learn, pymcdm, Scipy.
- Cloud platforms: Google Cloud, BigQuery.
- Frameworks: GitHub, GitLab.