Analyst
CurrentPredictive Analysis1. Crawled news articles about Facebook with Webhose API. 2. Applied text pre-processing and noise removal with Regex package and Natural Language Toolkit (NLTK) and deduplicated the data with LSH SimHash and Word2Vec.3. Calculated compound sentiment score (polarity score) of each article with NLTK Sentiment Intensity Analyzer. 4. Implemented LDA training and topic modeling with Gensim package and developed topic taxonomy based on LDA topic model.