Passionate about data,responsibility and perseverence my core work values, improving the efficiency of the way this world functions with my skills is my vision
-
Experience
I have completed my Bachelors in Computer Engineering from University of Mumbai and am currently pursuing Masters in Business Analytics from University of Maryland, College Park. During my undergraduate studies, I pursued internships that improved my coding skills such as Python, R and Tableau. I had the responsibilities of performing A/B Testing, Causal Inference and using descriptive and inferential statistical modelling to identify and analyse critical business metrics. Furthermore, having experience with constant team meetings and 1-on-1 mentorship from my supervisors, during my 1 year machine learning internship at Decure, I also learnt about experimental designing and test development of a business product. Coming to my Masters, I have over 6 months of machine learning research experience now through which I also learnt how experimental designing and scientific modelling works. I was mainly responsible for implementing multiple state-of-the-art ML and ensemble models, studying their differences and noting their accuracies of messy data cleaned by me. Apart from my internship experience, I have pursued and won multiple hackathons. In one of them, my team and I built an end-to-end website that predicted stock returns three years into the future and also supported chatbot capabilities. In another hackathon, my team and I built an end-to-end AWS product that scanned math problems and returned Google links as solutions. Recently, I participated in Deloitte’s COVID19 Analysis Workshop wherein I found some exciting insights like - states which had an overall decreasing no.of new daily COVID cases tweeted more positively than those which did not. Currently, as a Teaching Assistant for Financial Analytics, I am happy to share my communication, presentation and people skills have gotten a boost through this experience. Apart from academics, I absolutely enjoy reading about cellular food technology, neurology and playing badminton.
-
Projects
Projects
Airbnb High Booking Rate Prediction Using Machine Learning Techniques (R, xgboost, randomForest, caret)
● Led a team of five to victory; feature engineered 100 variables; performed Logistic Regression, K-Nearest Neighbors, Random Forest and XGBoost algorithms to predict Airbnb’s high booking rate with accuracy of 85.16%
Business Analytics Extract-Transform-Load (ETL) Suite (SQL, MySQL, Tableau, Python)
● Analyzed web scraped data using advanced SQL queries (MySQL) and Tableau dashboards
● Uncovered that Barbeque and teashops are the most well-received and affordable places amongst the public in College Park
Quora QnA Topic Modelling Using Latent Dirichlet Allocation (Python, Spacy, Pandas, NumPy, Sklearn)
● Fit and transformed a Quora QnA database into a TF-IDF vector; Used unsupervised methods of LDA and Non-Negative Matrix
Factorization (NMF) to assign topics to each question based on Spacy’s “en” library
Data Analytical Suite on YouTube Video Data (Seaborn, Matplotlib, WordCloud, TextBlob, Spacy)
● Performed data munging, cleaning, transformation, and visualization on YouTube video content
● Discovered that an Airbnb listing that was largely available for the next 30 days had a very high tendency of being available for the next 360 days which suggests that Airbnb earnings act as the bread-and-butter of many households
Hackathons/Workshops/Extracurriculars
Deloitte COVID19 Tweet Sentiment Analysis (Latent Dirichlet Allocation, K-Means Clustering, Sentiment Analysis) - Aug 2020
• Used tokenization, lemmatization, pivoting and melting, emoji cleaning techniques to clean a large dataset of 20
million Twitter tweets (March-April 2020)
• Performed Sentiment Analysis and Data Visualization on this data to analyze the US Government’s Response
towards the COVID pandemic.
• Discovered that states that had a constant number of daily active cases during this time tweeted more positively than
the ones which did not
Adobe Analytics 2020 Challenge (Adobe Sensei) - Upcoming - Sept - Oct 2020
AWS BlazeClan Hackathon (AWS Rekognition, Lex, S3, Glue, ELB, CloudWatch, Lambda) – Winner - 2018
• Hosted an end-to-end website on Amazon AWS services using AWS Rekognition and AWS Lex for scanning math
problems and providing Google links as solutions to the scanned problem; Used AWS Glue and Lambda to implement
a complete ETL Pipeline for data extraction, transformation and storage in AWS S3 bucket
Money Control Hackathon Challenge (Python, FBProphet, Django) – Runners Up - 2017
• Performed Time Series Forecasting on terabytes of stock data extracted from Quandl API through a period of 5 years and
Integrated with a Django website; achieved an accuracy of 88.65%
• Visualized forecasted stock price values through a per