Explored 1,882 charged-off bank accounts to evaluate the effectiveness of five debt recovery strategies based on expected recovery amounts. The dashboard provides insights into recovery performance, highlighting which strategies yield the highest returns and the impact of increased effort on recovered amounts. This analysis can help optimize future recovery efforts and resource allocation for maximum debt recovery efficiency.

Skills: Exploratory Data Analysis

Analyzed 471 passenger reviews collected from 2011 to 2024 to evaluate Kenya Airways' (KQ) customer satisfaction. The data, sourced from Skytrax, assessed multiple customer service aspects including value for money, inflight entertainment, seat comfort, and food quality. The goal was to identify pain points in customer experience and determine key areas of improvement for KQ.

Skills: Web Scrapping, Exploratory Data Analysis

Analyzed data from iFood's 2014 marketing campaign to identify patterns of customer behavior and attributes with regard to accepting or rejecting a gadget offer after a marketing campaign. This was in line with the company's objective of understanding their customers' profiles beside making a predictive model that would assist them in maximizing their profits in further campaigns.

Skills: Data Cleaning and Preprocessing, Data Visualization, EDA, Feature Engineering

Developed ML models to aid iFood’s marketing department in planning their marketing campaign effectively. The models correctly predicted over 75% of the customers accepting an offer and had an overall accuracy of more than 87%. Deployed the models in an application that supports both a single entry of customer information or a batch upload of various customer information in a CSV file. The application has an integrated Tableau dashboard in the single-entry interface to help make informed decisions based on the model's predictions.

Skills: Predictive Modeling, Hyperparameter Tuning, Handling Imbalanced Data, Model Deployment, Data Visualization.

Developed ML models with over 95% score in precision, recall and accuracy for crop selection based on essential environmental factors such as rainfall and temperature as well as soil measures: nitrogen, phosphorus, potassium, and pH levels. Deployed one of the models to a Streamlit app to facilitate user-friendly predictions and informed decision-making.

Skills: Feature Selection, Multi-Class Classification, Model Deployment