Analysis telecom data using Business Intelligence

Vijaya Krishna Yadlapalli

Master's thesis

Abstract:

Telecom companies have a competitive market as gaining new customers is very difficult compared to retaining an existing customers. As retaining advertisements can be used to prevent them, but we need some efficient solution to stop the churn. That’s why we need churn perdition and with this prediction, we can get an idea what new changes they need to do. In my thesis, I focused on IBM the telecom dataset. The data was containing the numerical and categorical features. The dataset has one column which is showing the customer is churning or not. The dataset is containing customer basic information, account information and services they are using. To analyse the data, I used python libraries: pandas, Numpy, and matplotlib. I applied data pre-processing, transformation, and cleaning on data. After analysing the data, I made hypotheses that tenure, monthlycharges, totalcharges and contracts are the main reason for churn. Then, I applied the machine learning algorithms like SVM, Random Forest, Logistic regression, and KNN and analysed the important features based on the model. Random Forest has supported my hypotheses, but SVM and Logistic Regression model don’t supported my hypotheses.

Keywords

business intelligence, customer analysis, data analysis, reporting, python

Jazyk práce: English

Datum vytvoření / odevzdání či podání práce: 1. 4. 2022

Obhajoba závěrečné práce

• Vedúci: doc. Ing. Jan Tyrychtr, Ph.D.

Czech University of Life Sciences Prague

Faculty of Economics and Management

Master programme / odbor:
Systems Engineering and Informatics / Informatics