Check out my favorite and most recent project- NortheastWS
A web app that mines 25 top news stories from twitter, groups them by similarity using an ensemble of ML models and informs you about the context of the news by studying the subjectivity of the tweet that links the story (official handle) and polarity of all the replies to said tweet.
- News paper like delivery of day's top stories.
- Prevents doom scrolling twitter for news.
1. Who are you?
I am a data scientist who is diving deeper into NLP by the day. Meanwhile, I am wrapping up a Master's in Operations Research, reading a new piece of fiction, and watching "Billions" again.
Recent Towards Data Science article on mining all replies to a tweet: bit.ly/TDS-tweet-replies
2. What do you know about Data Science?
I have worked with:
Programming Languages/Tools:
Python, SQL, R, Git, tensorflow, keras, huggingface, neuralcoref, sklearn, numpy, pandas, plotly, NLTK, spaCy, matplotlib, seaborn
Machine Learning Tools/Models:
Regression (Linear, Logistic, Ridge, LASSO), KNN, K-Means/Medoids, Naïve-Bayes (Multinomial), Dimensionality Reduction (PCA, t-SNE), SVM, Decision Trees (Random Forest, Bagging, Boosting), Deep Learning (CNNs, RNNs, LSTM), Topic Modeling (LDA, Gibbs Sampling), Named Entity Recognition
Deployment Tools :
Dash, Heroku, AWS (ElasticBeanstalk, S3)
Coursework:
Data Mining, Machine Learning, Design and Analysis of Algorithms, Stochastic Methods, Applied Stochastic Methods, Simulation and Stochastic Modeling, Data Analytics and Predictive Modeling, Decision Analysis, Linear Programming, Discrete Optimization
3. What does your experience look like?
I have internship experience as an Analyst with Immersed Games- a 43 North winning Ed-Tech Startup. I helped them mine insights about product engagement metrics and user behaviour and made web apps that quantify these using pandas, numpy, plotly, Dash and some SQL.
In a fellowship with SharpestMinds, where I built a news aggregator app that delivers additional information about News Articles such as the polarity of the discussion they evoke, the subjectivity of the official tweets that link the news article in twitter.
In my master's thesis, I study media polarity and hyperpartisanship using an NLP and Deep Learning.
4. Anything else?
Yes. I have co-authored two papers. One surveys the applications of blockchains in humanitarian operations (in revisions). The other one predicts fire risk using weather data across key US states using a Gradient Boosting Tree.
In a past life (out of several), I have created content for LaLiga (the soccer league in which Lionel Messi plays), NBA Basketball School, and Roland-Garros (the grand slam that Rafa Nadal has won over a dozen times).
I am also working towards building a fact based news commentary/op-ed platform called "Chutzpah". More Coming Soon.