Data Science Portfolio

---

 The Bank of England prudentially regulates and supervises financial services firms through the Prudential Regulation Authority (PRA).  This project aims to enhance the use of Quarterly Earnings Call data sets to improve their risk assessment of individual firms and, in doing so, maintain financial stability. 

• Topic Modelling
• Sentiment Analysis
• LangChain, RAG

Analyzed historical book sales data to make data-driven decisions about their future investment in new publications.

  Time Series Analysis and Forecasting with:

• ARIMA, SARIMA, SARIMAX
• Time Series Forecasting with machine learning and deep learning
• Gradient boosting models – Light GBM, XGBoost
• RNN, LSTM, GRU, CNN and Hybrid Techniques 

Analyzed customer feedbacks to understand what motivates members to join and what factors influence their behaviors once they have joined. 

  NLP and LLM:

• Text Classification / Sentiment Analysis / Topic Modelling / Text Summarization / Machine Translation
• LLM tiiuae/falcon-7b-instruct - Huggingface
• BERTopic
• LDAmodel from Gensim

Examined student data to predict whether a student will drop out and help institution's financial stability and students’ academic success and personal development. 

Supervised learning with:

• XGBoost
• Neural Networks - tensorflow, keras

Developed a robust customer segmentation to assist the e-commerce company in understanding and serving its customers better. This helped create a more customer-centric focus, improving marketing efficiency.

Unsupervised learning with:  

• Hierarchical clustering 
• k-means clustering

Develop a robust anomaly detection system to protect a company’s shipping fleet by evaluating engine functionality. 

Anomaly detection with:

• Interquartile Range (IQR)
• one-class SVM
• Isolation Forest 

Analysed how the number of visits impacts revenue generation on an e-commerce platform.

 A/B Test using PyMC linear regression model.