BACKGROUND
A company stores data about its processes and operating results (sales, leads, etc.). They also track customer satisfaction regarding their products on social media (mainly Twitter).
Problem
The company wants to analyse the correlation and influence that the comments of its customers and prospects on social media have on its business operations and economic results.
Benefit
Being aware in real time about the influence of customer satisfaction in daily operations is of an incredible strategic advantage (customer centricity).
METHODOLOGY & results
Architecture: Cloud architecture based Oracle Cloud Platform using Hive as a data storage system.
Developing language: Python/R.
ML techniques: NLP techniques and supervised learning algorithms (LDA-Linear Discriminant Analysis and Random Forest).
Results: The system is able to detect 90% of the sentiment of social media messages recovered, and to obtain the degree of influence on business results.
