Automated Text Analysis

20-23 september 2021

Elisa Volpi, CUSO

Nikolay Marinov, Houston, USA

Computational-Text-Analysis-for-PoliticalScience

Session 1: 

· Github 

· Jupyter Notebooks 

· Python Interactive Shell: Print Function 

· Assigning Variables and Values 

· Understanding Different Types of Objects in Python 

· Mathematical Operations 

· Boolean Operations 

· Transformation of Objects 

· Input function 

Session 2: 

· Understanding Python Strings 

· Combining Objects 

· Libraries in Python 

· Data Structures and Containers 

· List Comprehension 

Session 3: 

· Loops : for loops; if, else, elif loops - conditional loops; while loops; continue and break commands 

· Appending Lists 

· Defining Functions 

· Revise list comprehension 

· Join command 

· Opening, Closing, Updating text files: codecs library; Python inbuilt open command

· Basic Text Processing Operations: lookforwords; remove punctuation; remove digits 

Session 4: 

· Loading data into Python in different ways (Codecs, OS, etc.) 

· Manipulating data structures 

· Removing Stopwords

· Understanding and Building a Text-Cleaning Pipeline 

Session 5: 

· Managing Data with Pandas 

· Manipulating Data with Pandas 

· Mini Data Visualization Session 

· Short intro to Numpy 

Session 6: 

· Natural Language Processing Cleaning Pipeline 

· Adding to the NLP Pipeline: Tokenization Extra 

· Lambda Operator 

· More Visualization Methods 

Session 7: 

· Adding to the NLP Pipeline: Lemmatization 

· Adding to the NLP Pipeline: Stemming 

Session 8: 

· Adding to the NLP Pipeline: Pos-Tagging 

· Adding to the NLP Pipeline: Named Entity Recognition (NER) 

· Senses 

· Entity Linking 

Session 9: 

• Word-Embeddings and Vector Space Representation 

• Cosine-Similarity 

Session 10: 

Unsupervised Techniques 

· Clustering 

· Topic-Modelling

Session 11:

Supervised Machine Learning Techniques: 

• Classification: Naive Bayes; Random Forest; Logistic; K-Neighbors; Support Vector Machine 

• Evaluation Methods of Classification: Precision, Recall, F1 Score; Confusion Matrices; Cross-Validation 

Session 12:

• Different Ways to Approach A Sentiment Analysis Task: Based Sentiment Analysis; Classification; Rule

Session 13:

• Twitter Scraping and Data Cleaning Mini Session: 

Schedule:

Monday September 20th: 10:00-12:00 and 13:00-15:00

Tuesday September 21st: 10:00-12:00 and 13:00-15:00

Wednesday September 22nd: 10:00-12:00 and 13:00-16:00

Thursday Septmber 23rd: 10:00-12:00 and 13:00-16:00

ZOOM !

20