Recent Projects

Projects

1. Image Forgery Detection: In the digital age, the authenticity of images plays a crucial role across various industries, from journalism to identity verification. However, the rise of advanced image manipulation tools has made it challenging to verify the genuineness of these images. This project focuses on detecting tampered images using machine learning and image processing techniques. The project consists of the following four sub-projects:

   1. Document Forgery Detection With SIFT : Passive image forensics using the Scale-Invariant Feature Transform (SIFT) algorithm and homography estimation. Learn more

   2. Image Forensics with TensorFlow : In this project, we used googles deep learning algorithm (TensorFlow) to test if images have been digitally manipulated or not. Learn more

   3. Image Forensics with Deep Learning (Pytorch): Utilizes Convolutional Neural Networks (CNN) for image forgery detection. Learn more

   4. Decoding Encoded Census Data Using Pyspark: Analysis of data from the United States Census Bureau with the aim of decoding the data that was encoded for security and privacy compliance. Learn more

Completed Projects

Here are a few of the projects that we have completed.

1. Energy Generation, Supply and Consumption Visualization: We used PowerBI to connect to data sources related to Eskom’s power generation and created visuals to derive the insight required to appropriately answer several critical questions about the SA Energy crisis. Learn More Note that you must have a valid microsoft account to view this project

2. Scrapping S&P Historic Stock Price Data from the Internet: We used the yfinance library and python to scrape stock data from the S&P 500 (Standard & Poor's 500). Learn More.

3. Mortgage Amortization Project: We compared three different mortgage loans, each with a different interest rate, term, and amortization schedule. Learn more

4. Load shortfall Prediction: Predictive modelling of load shortfall between energy generated by means of fossil fuels and various renewable sources was built. Learn more

5. Stock Inventory prediction: Prediction of the right quantity of the products to be stocked for sale by a store that serves several countries from a common stock inventory to prevent overstocking or understocking. Learn more.

6. Asset trading Model: This trading model uses the 200day moving average to determine when to buy or sell an asset. this example trades crypto but the model can be adapted for stocks and other assets. Learn more

7. Basic Crypto Trading Model: Enforcing transaction protocols to buy or sell crypto assets when certain market conditions are met. Learn more

8. Identifying and Applying Risk Metrics Associated With Financial markets: Modelling returns of different asset classes. We observe custom volatility metrics, along with how to model them in Python. Learn more

9. Using Statistical Distributions to Model Asset Returns: Evaluation of distribution of asset returns for normality or otherwise in order to correctly model risk and returns on assets. Learn More

Other Projects

Here is a list of even more projects that we have completed for our clients some of whose identity we are not allowed to disclose due to NDAs to which we are signed.

1. Anomaly detection for a financial institution: This project involves the detection of anomalous transactions. It is useful in the early detection of frauds and abuse of bank processes, allowing the institution to take action to prevent loss of funds or other security breaches.

2. Cloud Migration: We migrated the entire operation of our client from a completely onsite operation to a cloud based operation with full security implementation of the least privilege protocol.

3. Churn Prediction: We generated a model that predicts churn by analyzing customer data to enable the classification of customers into groups that are likely to churn based on known features. The company can take pre-emptive steps based on these Predictions to prevent churn by targeting remedial efforts to these customers.

4. Customer classification: Classification of customers into categories based of behaviors, demographics etc. to allow for targeted marketing. This particular project resulted in a 32% improvement in the rate of conversion for our client.

5. Anomaly detection for electricity consumers: In this project, we detected anomalous patterns in the consumption of electricity. This allowed us to be able to detect electricity theft and by pass of meters in real time. Our client previously relied on manual detection by field staff which was both labor and capital intensive. Our solution resulted in an 87% detection of electricity theft wit only a 2% case of false positives, greatly improving our clients bottom-line.

6. Personalized Chat Bot: We leveraged on open AI GPT 4 model to create a transformer model that was deployed as a chat bot, trained on our clients information. This bot was able to provide visitors to our clients site with accurate and up to date information on their processes on the spot saving the client time and other resources.

7. Project Portfolio Management Project for a construction company: The Project Portfolio Management" database project is aimed to streamline and enhance the operations of PSP Alpha, a construction and property development company. The comprehensive system we developed was an efficient project tracking, resource management, quality assurance, financial oversight, and client/stakeholder engagement system. It centralized document management, timeline tracking, risk assessment, and reporting.

8. Youth engage Plus: The YouthEngage Plus project is a multifaceted database management system tailored to the needs of an organization dedicated to youth development and community engagement in Malawi. This all-in-one solution seamlessly integrates features for member and volunteer management, donor and fundraising tracking, event and activity registration, program evaluation and impact assessment, resource and inventory management, communication and outreach, and document / content management. By providing a centralized platform, YouthEngage Plus empowers our client to efficiently manage its operations, measure program effectiveness, strengthen stakeholder relationships, and streamline organizational processes, all while ensuring data security and privacy.

9. Digitization of Legacy project and thesis documents and creation of project repository and Database for a Tertiary Institution: We took thousands of old students' projects that were only available in hard copies, created digital copies of these projects and integrated them into their existing students repository. Because of our work, these materials which were hitherto unavailable to students as there were limited copies of them are now digitally available to potentially thousands of students at once. We also implemented access restrictions to the repository in line with the instruction of the client and provided an option for the monetization of the digital asset that we created.