Now Online: 2024 MITC Big Data Science Capstone Project Exhibition
Dear community (university and beyond)
We are thrilled to share the incredible work showcased at the 2024 MIT808: Big Data Science Capstone Exhibition at the University of Pretoria. This year, talented students formed 11 teams to tackle diverse and real-world problems through innovative data science solutions. This is not just a showcase for the MITC program, Data Science or Artificial Intelligence, but also showcases transdisciplinary research at the University of Pretoria. We had multiple UP departments, faculties leading the projects with a few external project leads from other universities/organisations as well. The physical exhibition was held on June 7 and supported by a gift from Google Tensorflow.ย
The online exhibition is now available. 2024 projects are viewable (videos, abstracts and poster) hereย https://up-mitc-ds.github.io/808exhibition/papers.html?filter=year&search=2024
Here are some highlights:
1. Measuring the Societal Impact of UP Research
Students developed an end-to-end solution to analyze and visualize research data from the UPSpace online repository. This tool helps the UP community assess research output, identify trends, and suggest potential transdisciplinary collaborations.
2. Detecting Fungal Diseases in Trees
Using computer vision and machine learning, this project automated the detection of fungal diseases in Eucalyptus trees, reducing the labor-intensive task of manual detection and significantly improving accuracy.
3. Locating Habitat for Endangered Cape Parrots
This project leveraged RGB image analysis and logistic regression to identify yellowwood trees in dense forests, crucial for the survival of the endangered Cape Parrot.
4. Bee Species Identification
Supporting research by the Social Insects Research Group, students used the YOLOv8 CNN model to classify different bee species, providing valuable insights into insect biodiversity.
5. Multimodal Sentiment Analysis of Afrikaans Musical Videos
Addressing the lack of informed consumption of Afrikaans music videos on social media, this project developed a recommendation system using sentiment analysis to enhance user experience.
6. Closing the Gender Gap in Media
Using NLP, students analyzed South African media to assess the representation of women, LGBTIAQ+ community members, and people with disabilities, promoting diverse voices in the media.
7. Predicting South African Electoral Results
A robust predictive model was developed to forecast election outcomes using historical data, aiding strategic planning and decision-making in political campaigns.
8. Cacao Plant Leaf Disease Detection
This project applied deep learning techniques to detect and classify diseases in cacao plants, supporting Ghanaian farmers in improving crop resilience and productivity.
9. Automating Rapid Elephant Population Assessment
Students used YOLOv8 to automate the assessment of elephant populations from aerial images, enhancing conservation efforts with accurate and efficient data.
10. Maize Disease Detection
An app was developed using deep learning models to provide real-time diagnostic feedback for maize diseases, helping West African farmers safeguard their crops.
11. Automated Rapid Elephant Population Assessment from Aerial Photos
This project created computational tools to streamline the estimation of key population parameters of African Savannah Elephants, supporting effective conservation strategies.
These projects demonstrate the power of data science in addressing real-world challenges. We are incredibly proud of our students and their contributions to these impactful projects. We are grateful to all project leads.
For more details on these projects and to see the creativity and skills of our students in action, please visit our online exhibition.ย ย https://up-mitc-ds.github.io/808exhibition/papers.html?filter=year&search=2024