Private Practice

House Price Prediction

Lagos, Africa's most populated city is home to over one million houses currently occupied. However, there seems to be at least half a million houses up for sale.
Most times, people's choice of house location is based on the price of the house. This project aimed to provide a means to predict house price per location.
Python's BeautifulSoup package was used to scrap data of houses available for rent in Lagos from one of Lagos' largest real estate's websites, Property Pro.

Regression analysis was performed on the output csv file to create a prediction model that yielded about 95% accuracy.

Tool Used: Python.

Classification in R

In this task, 8 different classification models were compared to predict if a customer will be given a loan using R programming language.
The ouput was visualised to help the user compare these models. .
The details of this task are published in a Medium article.

Tool Used: R Programming Language.

Association Rules Mining in R

Association rules mining is also known as Market basket analysis.
It is one of the most common and efficient methods used by companies to predict what products customers could possibly buy together.
In this task, I have deployed this method using R programming language to determine what clothing items customers are likely to buy together from an online store. Further detail into the task is available in my Medium article.

Tool Used: R Programming Language.

Loan Offering Prediction

In finance, a loan is the lending of money by one or more individuals, organizations, or other entities to other individuals, organizations, etc. The recipient (i.e. the borrower) incurs a debt, and is usually liable to pay interest on that debt until it is repaid, and also to repay the principal amount borrowed.
The whole process of ascertaining if a person will be offered loan or not might be tedious hence the need to automate the procedure using selected machine learning models.

An explorative, descriptive and predictive analysis is carried out on the dataset and each observations were recorded and meaningful insights were gain from them.
Six different classification models were used for this project; Logistic Regression, K-Nearest Neighbours, Random Forest Classifier, Decision trees, Gaussian Naive Bayes algorithm and the Support vector classifier.
The best performing models after hyperparameter tuning were Logistic regression, Random forest and Support vector classifier.
A RMSE value of 0.44 was recorded and an accuracy score of about 80% was achieved.

Tool Used: Python.

African Economic Crisis Analysis

This project was aimed at analysing economic progress and decline in Africa between 1860 and 2014.

Data from different African regions was gathered and analysed to create insights on economic behavior in the time period.

The following conclusions were gathered:

1. Central African Republic have the highest occurence of systemic crisis followed by Zimbabwe and Kenya.
2. Systemic crisis has impact on banking crisis. This is shown by overlapping of both crisis in visualization.
3. There was a general increase in exchange rates from 1990.
4. Zimbabwe defaulted both in external and domestic debt. However, this default was not dependent on if they experienced bank crisis that year.
5. Most countries experienced currency crisis before experiencing banking crisis
6. Countries had an inflation crisis despite not being hit by a banking crisis.
7. A rise in annual CPI co-incides with the same time as bank crisis (and systemic crisis) which usually co-incides with inflation crisis.

Tool Used: Python.

Other Projects

I have worked on a wide variety of projects, more of which can be found on my GitHub. Examples include:

1. Sentiment Analysis in R.
2. PySpark DataFrame and RDD Analysis.
3. HiveQL Data Analysis Implementation.
4. Breast Cancer Prediction.
5. Wine Quality Prediction

Tools Used: HiveQL, Pyspark, Python, R.