LATEST PROJECTS
Project | 01
PREDICTING DELIVERY TIME.
In this Zindi competition, I build a model that predicts an accurate delivery time, from picking up a package to arriving at the final destination. An accurate arrival time prediction will help all businesses to improve their logistics and communicate an accurate time to their customers. I used Simple Linear Regression, Multiple Linear Regression, Lasso, Ridge and ElasticNet.
Project | 02
PREDICTING EMPLOYEE PROMOTION.
In this Project I build a Machine Learning model that will help the owner of a company to detect features which will help him decide on employee promotion. Logistic Regression, Decision Tree Classifier, Random Forest Classifier, Support Vector Machine Classifier and Xgboost Classifier were developed. Maximum Hard Voting Ensemble is used to provide an optimal model with 80% accuracy.
Project | 03
STROKE PREDICTION IN PATIENTS.
This work was to bring forth the factors influencing stroke. As such, the approach chosen was to develop a predictive model using the provided train and test dataset. The programming or statistical package employed here was Python using data wrangling and visualization libraries Numpy, Pandas, Seaborn and Matplotlib.pyplot. In modeling, classifiers chosen were Decision Tree, Random Forest, Logistics Regression and XGBoost. Finally, the Maximum Hard Voting Ensemble Model was chosen to encompass base classifiers with 98% accuracy. Feature importance was used alongside to visualize pressing factors of stroke for each base classifier.
Project | 04
PREDICTING CUSTOMER CHURN IN TELECOMMUNICATION.
This project is a Classification Problem to predict customer churn in telecommunication. Exploratory Data Analysis, Data Processing, Feature Engineering, Imputing Missing Data, Smote for Imbalance Data target feature are all employed. Classification models used are XGBoost, Logistics Regression Classifier, Naive Bayes, Decision Tree, Random Forest, K-Nearest Neighbors, Support Vector Machine. These features are used as base classifiers under Maximum Hard Voting Ensemble Model. An optimal model is obtained with 77% accuracy.
Project | 05
MACHINE LEARNING MODEL TO PREDICT FACEBOOK ENGAGEMENT.
Technological advancement has led to the rise of social media. And as such section of the Ghanaian populace being the teen and adult have an account on Facebook, there being their source of news feed. Politicians are hence capitalizing on this development by hosting an official page to communicate their campaign and message to this section of voters. It is then crucial to time and determine the content being put on their page to spark proper engagement (Likes, Shares, Comments). This projects seeks to build a machine learning model to predict engagement level.
Project | 06
BANK CUSTOMER SEGMENTATION.
Having a Bank data on their transactions, I built Unsupervised model using Hierarchical Clustering and K-Means Clustering method to segment customers. Three (3) clusters were settled on by age against length(time with company) of the customers. This results goes a long way to strategically plan marketing campaigns to reach desired customers.
Project | 07
DASHBOARD FOR INSURANCE COMPANIES IN GHANA
In this freelance project, data was sourced from National Insurance Commission (NIC) in Ghana, that publish quarterly data on insurance companies performance. Here a dashboard was created to give the public, insurance companies and stakeholders knowledge in the insurance industry. Especially on market share, complaint by policy holders, assets, claims among others.