Camila Matoba

Data Scientist - Text analytics

09.2022 - 02.2023

EY, Consultancy - Amsterdam, The Netherlands

Core Skills: Customer Insights (Python), Databricks, Natural Language Processing (GPT, BERT, T5), Deep Learning (TensorFlow + CUDA)

A company in the healthcare sector implemented a new robot to help in daily tasks with their elderly clients. The robot is important as their staffing was overworked as it is hard to find new workers in their sector. To evaluate how their clients approved of the robot, they gathered feedback in the form of multiple interviews over time. As the interviews content is open text, analysing manually would demand even more work from their staff.

I used natural language processing (NLP) techniques (Python and databricks), in particular emotion detection and aspect extraction, to automate the analysis of the data. The analysis helped their staff by deploying a pipeline that automatically cleans the data, extracts the most important aspects and their respective sentiment involved. In another shorter project, I investigated consumer reviews for detecting what products led to which aspects to their branding.

The final report gave an overview of most recurrent aspects of the robot that their clients liked or disliked, so that they could derive data driven actions out of the insights without the need to do it manually. Furthermore, Guidance on how to make these questionnaires more efficient were also promoted, as the framework proposed can be used in their future projects.