Eugene Ambe Ndamukong's repositories
Search-for-clients-with-high-propensity-to-own-loan-mutual-fund-and-credit-card
Machine learning algorithms were used to select clients with a high propensity to own loan, credit card and mutual fund. These clients are expected to maximize the revenue of the bank.
NLP-and-machine-learning-on-text-data
Building machine learning model to classify doodle images
Predict-unhealth-healthy-cassava-leaves-using-Deep-learning
This involved using deep learning models to predict the health status of cassava leaves
Churn-prediction-for-Telco
Involved using machine learning classification models to determine if subscribers to the company will churn or not. Furthermore, K-means clustering was used to determine the retention campaign for customers with a high chance of churning
Natural-language-processing-tools
Collection of NLP tools for mining social media data. Developed for a start-up company
terminal_blog
Program to introduce MongoDB and Object-Oriented Programming by creating a simple terminal-based blog
Capstone-project-Advance-data-science-specialisation-
As the final project for the IBM certificate, I applied machine learning and apache spark tools to estimate a model that predicts recession in African countries
nlp-in-python-tutorial
comparing stand up comedians using natural language processing
First-Cameroon-project
The is my first data science project on Cameroon's development. It involved web scrapping and data visualization
Capstone-project-Battle-of-the-neighborhoods
This is part of the IBM data science professional certification. This involves analysis of location data
Clustering-and-Segmentation-of-Toronto-neighborhoods
This is part of the IBM data science professional certification. It involves K-means clustering of neighborhoods in Toronto
IBM-training-on-Classification-models-and-model-evaluation
This was a project done for the certification of "IBM data science professional". It involved data preprocessing, model building and model evaluation. The goal was to find a good model for predicting load payment
IBM-training-on-data-visualization
This part of the training for "IBM Data Science professional certificate". There two parts : The first part is investigating the distribution of a survey of data science fields and the second part is visual display of crime rate in San Francisco using a map
Hangman
This was a programming challenge made for 3D Hubs company based in Amsterdam, Netherlands. Hangman game(HANGMAN.ipynb file) played using 5 different letters of the clients choice. The client can only win if none of his letters match with letters of a random unknown word. This miss-match gives the client a rank of 1 (1st position). If the rank is higher than 1, the client looses because his/her letters matched at least once. Enjoy!
semiparametric-models-for-online-learning-data
The aim of this master thesis project was to evaluate the performance generalized additive models on online learning data. There are 2 r-codes to generate abilities for each model cases (GenData_ses.R and GenData2.R). Also, there were 2 files to analyze data for the 2 simulated cases (Analysis of simulated data.R and Analysis of simulated data 2_1.R). The real-life data was analyzed using "Analysis of real-life data.R"
price-of-chair-new
Revamped final project for our Complete Python Web course
Project-as-statistical-consultant
School project solved as a statistical consultant. "Project codes 2.sas" contains codes for a clustering analysis while "Retake project.R" contains codes to build a predictive model. The "Statistical_Consulting_Report_2.pdf" is the report describing the problem being solved and how it was solved. These results were presented in a seminar in the midst of 2 statistical consultants.
Affect-of-work-load-and-salary-on-job-satisfaction-
This project involved interpreting how work load and salary could affect the job satisfaction. A 2-way ANOVA model was used to understand this relationship.
Modelling-bodyweight-of-chickens
This was a project I did for a company Belgium. It was requested to extract useful information from a data set containing information about chickens. There was information about the body weight of chicken which I thought was commercially favorable for business. The bodyweight was regressed on other variables present in the data set
Prediction-of-loan-default
This analysis involved search for a model that was capable of predicting if a client will default or pay a loan. The analysis began with cleaning of the data to prepare them for modelling. Models used in this case included: Random forest, Support vector machine and Logistic regression. The best model was chosen based on performance indices like accuracy, recall, AUC.
Time-Series-analysis-of-Wages-in-the-UK-1855---1987-
The project involved searching an appropriate univariate model that was able to understand the change in wages over time and forcast the wages beyond 1987. Additionally, the effect of wage on employment was also investigated through multivariate models
Streaming-with-Spark
Source code(Task-3-1-5.ipynb) for streaming with spark along with text mining. Newspaper articles streamed from a server and their respective titles and description extracted. The extracted titles and description used in training a model offline. The model was then used to predict categories of newspapers while streaming
Creation-of-dashboard-for-visualisations-of-Amusement-park
This project involved investigating activities at an Amusement Park and a crime that occurred in 2014. Running the code Data visualization for communication and movement.ipynb generates a dashboard which is explained in ActivePresenter for Exam project.mp4.
web_blog
Simple web-based blog to introduce Flask, HTML, CSS, Bootstrap, and Jinja2.