NLP, Text Mining,
Data Science; Python, NLTK
One of the best-known Central European e-commerce merchants (under NDA) was looking to create a tool intended to analyze the customer’s feedback on goods purchased through their web marketplace. The business goals were to increase customer loyalty, drive business changes, and deliver an appropriate return on investment. Customer experiences fall into three basic categories: positive, negative or neutral. The goal was to perform in-depth sentiment analysis and detect the tone of voice for every single customer’s comment/review/post in social media pertaining to the purchased product.
The Client hired us for building their Dedicated AI and Data Science Team in our Offshore Development Center in Kyiv, Ukraine.
After data cleaning and munging, we tokenized every single word in customer feedbacks. After this, an Natural Language Toolkit was used to define synonyms, semantics, and overall tone of voice of feedbacks. Having aligned them with scores provided by these particular feedback authors, a section of manual business logic was brought in: language specifics, abbreviation, collocations, and vernacular expression played a significant role in the overall semantic analysis. Alongside NLP and semantic analysis, Data Science techniques were applied: based on the data from the social network that each customer logged in, a set of demographic features was defined, leading to a complex analytical solution.
We have set up a dedicated team comprised of highly-skilled analysts, mathematicians and software developers.
This solution helped the Client to define and upgrade their marketing and sales strategy, which resulted in 30% revenue increase within one year after the deployment.
The Client noted the overall high level of delivery and solution architecture.