About
Business Intelligence & Data Scientist
- Phone: +1 585-978-0434
- City: Rochester, NY
- Degree: Masters in Business Analytics
- Email: [email protected]
Skills
R 100%
Python 90%
Tableau 80%
Database 80%
Machine Learning 90%
Cloud Platform 70%
Resume
- A Data Scientist with a knack for deciphering challenging problems, with experience with large-scale data sets and telling stories using data.
- Seeking to leverage experience with quantitative analysis and statistical data mining to solve complex business problem using various data science technologies
Sumary
Tanay Bhardwaj
Driven and passionate data scientist, with strong technical skill and business acumen. Looking for new challenging problem to work on.
- Rochester, NY
- (585)-978-0434
- [email protected]
Education
Master of Science, Business Analytics
2018 - 2020
Rochester Institute of Technology, Rochester, NY
Relevant Courses : Principles of Statistical Data Mining, Time Series Analysis, Marketing Analysis, Database Management and Analysis, Introduction to Data Analysis and Business Intelligence, Financial Analysis
Bachelor of Technology, Electrical Engineering
2014 - 2018
SRM University, Chennai, India
Professional Experience
Graduate Research Assistant
2019 - 2020
Saunders College of Business, Rochester, NY
- Improved decision-making process by 30% by designing dashboard to provide a comprehensive snapshot of key performance metrics to cross-functional teams
- Reduced marketing expenditure by 15% YoY by analyzing historical admissions, social media data and identifying data patterns using Tableau
- Provided business insight to the marketing team by performing data analysis to support key decision making by leadership for the admission process and social media presence
- Advanced Tech Skills & tools: Tableau, Tableau Prep, Excel
Business Intelligence Analyst
2019 - 2020
Arcadia LLC, Rochester, NY
- Developed multiple SaaS Application worth $10000+ including text and sentiment analysis to enable small business based in Rochester to better understand customers
- Implemented better visualization with 5x faster load time by optimizing SQL query requests and designed a robust data collection process which improved leads generation by 10%
- Advanced Tech Skills & tools: R, Shiny, Python, PHP, SQL
Portfolio
- Forecasted less usage of Yellow Taxi in the year 2016 by at least 15% compared to 2015.
- Analyzed 200 GB of New York City Yellow Taxi Data consisting of 1.1 billion rows using R, Python and SQL
- Implemented .hdf5 file system with Vaex to reduce data read time by 80%, and Uber H3 library for Geo-spatial Data Visualization
- Developed ARIMA and Trigonometric seasonality, Box-Cox transformation, ARMA errors, Trend and Seasonal components (TBAT) Model and predicted ride frequency for the year 2016
- Conducted feature engineering, data cleaning, data visualization of high dimension, highly imbalanced dataset.
- Utilized google compute engine, h2o driverless AI on docker container to reduce the running time by 60%.
- Tested logistic regression, random forest, gradient boosting, deep learning models with random search.
- Classified predictor variable with an AUC value of 0.75, used imbalanced sampling methods for better AUC.
- Finished top 38% in M5 Forecasting accuracy Competition with score of less than 1 on kaggle.com leaderboard.
- Proposed the use of hierarchical time series method to a team of 5 people, which in turn decreased the RMSE value, and ended up increasing the ranking by 57% on kaggle.com leaderboard
- Performed exploratory analysis and extracted key metrics from the data.
- Implemented hierarchical time series analysis techniques to forecast the sales for the next 28 days at 10 stores in 3 different states of Walmart
- Presented the results to a group of 30 people. Leveraged 200,000 tweets to develop sentiment analysis model, ran hypothesis test to determine which company has better public sentiment.
- Performed Dimensionality reduction of text corpus using t-distributed stochastic neighbor embedding to decompose data to 2-dimension
- Built topic models using Tf*Idf, Covariates and latent Dirichlet allocation (LDA)
- Classified tweets into Positive, Negative and Neutral and Applied Naive Bayes, SVM, KNN for classification of the text. Used Bootstrapping and Cross Validation to achieve better accuracy.
- Obtained and cleaned global life expectancy data from kaggle.com; removed missing/incorrect data.
- Executed descriptive statistical analysis on all the variables.
- Selected the independent variables for regression analysis.
- Visualized the data using Tableau.
- Identified African regions with lack of factors that lead to global life expectancy.
- Suggested actionable recommendations to improve life expectancy in selected regions.
- Based on article in The Economist magazine about the death of Globalization and the need for a newly invented world order.
- Created own dataset for analysis and came up with model based on the article
- Used R to gather twitter sentiment in different part of world, and finding showed that the narrative is different in different parts of world.
- In countries like United States, the most frequent words were “Peace” and “hope” when compared to “Military” and “Hostage” in Near East region
- Visualized the data using Tableau, Performed Statistical analysis and Concluded that the author’s argument that globalization is dead seems to be true.
- Analyzed the USDA obesity data, determined which variables to include in the model.
- Performed data cleaning, exploratory data analysis to find insights from the data.
- Identified the most important factor affecting the obesity rate. Created a regression model for obesity rate.
- Performed t-tests and ANOVA to statistically confirm the findings.
- Provided actionable recommendations to lower the American obesity rate.
- Plotted and estimated Kaplan-Meier curve for e-commerce business data to visualize the probability of customer loss at a given point in time.
- Used Cox PH model to calculate hazard ratio and evaluate simultaneously the effect of several factors on customer survival.
- Performed Market Basket Analysis on insta-cart grocery data.
- Used to find the relationship between sets of food ordered together at once with each purchase.
