Data Projects
Data analytics projects — still learning, still building.
Digit Recognizer (Kaggle MNIST)
Designed and trained three progressively deeper CNNs from scratch in Keras/TensorFlow — stacking Conv2D blocks with batch normalization and dropout, tuning learning rate schedules with EarlyStopping and ReduceLROnPlateau, and diagnosing failures with a full misclassification breakdown. Landed at 99.34% test accuracy, missing just 66 out of 10,000 handwritten digits.
PythonTensorFlow / KerasCNN architecture designBatch norm & dropout tuningModel evaluation
What's next
- Build an interactive demo — draw a digit, get a live prediction
- Test it against messier real-world handwriting, not just clean/centered MNIST digits