This is a case study that will explore a problem related to school district budgeting. This problem is part of a competition posted on 'drivendata.org'. The training and test datasets can be found as separate attachments. Problem summary and the steps to build the classifier are all mentioned in detail in the jupyter notebook.
Here, we use Jupyter Notebook to build a multi-class-multi-label classifier in Python to automatically classify items in a school's budget which makes it easier and faster for schools to compare their spending with other schools. Pipeline, FunctionTransformer, FeatureUnion and CountVectorizer were some of the tools used for training and testing the model for accuracy and the metric used here is 'log loss'.