The document discusses developing a classifier to identify leaky and non-leaky data in Chicago city council daily purchase logs. It describes collecting a data corpus from public records, developing an ontology to classify data points, and using techniques like neural networks, SVM, decision trees to build classifiers. The best performing model is a decision tree classifier that achieves 84.45% accuracy and F1 score of 0.9038878 on test data. The deliverables will be a data classifier and evaluation of model performance against different classifiers.