Summary of "Get to know me | Tatenda Matika | Zimbabwe - Africans in Data Science and AI"

Scientific concepts, discoveries, or nature phenomena mentioned

Data science as a discipline
- Combines domain expertise, programming, and mathematical/statistical skills to analyze data and generate insights, including predictions/forecasts and classifications.
Machine learning (ML) connection
- Data science “trickles into” machine learning, used for tasks like prediction and classification.
Mathematical science for data science
- The training pipeline is described as moving from mathematical science into data science/ML, emphasizing modeling and pre-processing (with comparatively less mention of exploratory data analysis).
Data analytics for business decision-making
- Applying data analytics to business organizations, referenced through “business and decision making using data.”
Neural networks for audio classification
- Project concept: improving an audio classification approach by converting:
  - Artificial Neural Network (ANN) approach → Convolutional Neural Network (CNN) approach.
- Uses signal processing:
  - Convert audio waveforms into spectrograms (described as “picture representations” of audio).
  - Feed spectrogram “images” into a CNN for classification.
Forecasting and market analytics
- Current work concept: market analytics and price forecasting.
- Mentions implementing forecasting models via a web application.
Literature review as part of modeling workflows
- Daily workflow includes researching and performing a literature review to support forecasting implementation.

Methodology / workflow outlined (as described in the video)

Audio classification pipeline
1. Start with existing audio-classification code using an ANN.
2. Modify it to use a CNN.
3. Transform audio files into spectrograms using signal processing.
4. Train/use a CNN to perform audio classification.
Typical day / work sequence for analytics & forecasting
- If data is available:
  - Open and inspect data in Excel for a rough understanding.
  - Perform analysis using programming (Python/R).
  - Proceed to forecasting.
- For longer-running projects:
  - Alternate between literature review/background research and implementation.
  - Implement forecasting models through a web application.
  - Perform analysis as needed.

Tools and platforms referenced (supporting the above concepts)

Programming/data tools: Python, R
Data/business tooling: Microsoft Power BI, Excel
Statistical software: IBM SPSS
Code-sharing/learning mentioned: GitHub, tutorials, LinkedIn, a personal website