CUERIS

SENTIMENT ANALYSIS OF CUSTOMER REVIEWS

SENTIMENT ANALYSIS OF CUSTOMER REVIEWS

Business goal:

To conduct an analysis of client reviews to gauge satisfaction with an e-commerce platform.

For more information, contact us

Challenge:

The opinion or “sentiment” data, generated through social channels in the form of reviews often includes comments that can be invaluable for businesses looking to improve products and services, make more informed decisions, and better promote their brands. The key to business success with sentiment data lies in the ability to mine vast stores of unstructured data for actionable insights requiring sophisticated tools such as Natural Language Processing (NLP). The challenge here was to gather, store and analyze plain text and automate the workflow to process large amounts of text to predict sentiment.

 
 

Solution:

Algorithms that can automatically collect customer reviews from webpages and store them were developed. Stored text was then structured according to different word and syntax characteristics (e.g. labelling adjectives with an underlying feeling) in order to analyze the sentiment of a reviewer. Machine learning techniques were used to make the algorithm more intelligent by feeding more data into the system.