Data Science tasks

This repository contains notebooks and code for three tasks:

1. Text classification: given a file with articles' headlines, classify if they are sarcastic or not. Techniques tried:

• TF-IDF and Logistic Regression
• word counts and Naive Bayes
• deep neural network using LSTM and pretrained Glove embeddings

2. Tabular data exploration, cleansing and classification:

• analyze corrupted dataset
• perform necessary cleaning
• visualize the data
• apply a classifier

3. SQL query: given three tables with students' data, their marks and classes taught at the university, write a simple query to find students that scored well on algebra classes.