Data Analyst

***, ******** ******, ******, **-***** Contact: **0 - 298 - 7468 Email: *****@****.***

https://www.linkedin.com/in/sharmilinag https://github.com/Sharmilinag Available: Immediately Education

New Jersey Institute of Technology – Newark, NJ

Master of Science in Information Systems, GPA: 3.5/4.0 Dec 2017 Related Courses: Data Science, Web Mining, Data Analytics, Cognitive Computing, Applied Statistics Future Institute of Engineering & Management – Kolkata, India Bachelor of Engineering in Electronics & Communication, GPA: 3.8/4.0 Jun 2013 Related Courses: Data Structure, Database Management System, Software Engineering, Algorithms Technical Knowledge

Languages Python, SQL, R

Machine Learning Frameworks Keras, Tensorflow, numpy, pandas, scipy, sklearn, NLTK, Amazon AWS Data Analytics Statistical Modelling, A/B testing, Hypothesis Testing, Data Mining, Predictive Modelling Data Visualization Tools Tableau, d3.js, c3.js, chart.js, leaflet.js, Python matplotlib Databases MongoDB, Oracle, DB2, SQL Server, MySQL, DynamoDB, MS Access Tools and IDEs QlinkView, Tableau, SAS, RapidMiner, Minitab, MS Office, Eclipse, PyCharm, Matlab AWS Technologies DynamoDB, Lambda, API Gateway, IAM Roles, Cognito, Amazon RDS, EC2 Work Experience

New Jersey Institute of Technology, New Jersey, USA Jan 2017 - Aug 2017 Research Assistant

Analyzed and Visualized Mel-frequency cepstral coefficients (MFCC) for recorded speech samples

Resolved critical issues for Feature Extraction using various ASR specific smoothing and imputation techniques

Technologies: Python, MFCC, Smoothing Techniques, Feature Extraction, Automatic Speech Recognition IBM, Hyderabad, India (Client: Westpac Bank) Dec 2013 - Jul 2016 Data Analyst

Implemented co-relation model to predict customer churn in a 30 days’ time window in Agile Project Methodology

Designed recommendation system to predict best credit card deals for customers using Market Basket Analysis

Developed Teradata Scripts to perform ETL on data received from heterogeneous sources like Mainframe files, APIs, XMLs

Created a distributed ETL pipeline that loaded new data to DWH 20% faster than the existing pipeline for the same task

Built complex SQL query framework to automate the generated scripts process for stored procedures

Designed tableau dashboard to represent analytics for clients’ data using complex aggregate and analytical functions

Technologies: ETL, SQL, Tableau, Business Intelligence, Financial Data Analysis, Market Basket Analysis, Agile Ardent Collaborations, Kolkata, India Oct 2012 - Nov 2013 Research Analyst

Built macro scripts and VLOOKUP in Excel and analyzed data with pivot tables and validated claim using hypothesis testing

Technologies: MS Excel, VLOOKUP, Hypothesis Testing, Pivot Tables, Large Scale Data Analysis Academic Projects

Automatic and quick detection of the genre of audio/music files (2017) Link

Implement regularized logistic regression to predict genres of music based on the MFCCs computed on the audio tracks

Achieved 91.2% F1-score and average prediction time of 58ms per track with a test set of 5000 audio files

Technologies: Python, sklearn, glob, numpy, scipy, pandas, scikits.talkbox, matplotlib, ROC Classification-Based Destination Recommendation System of Foursquare NYC dataset (2017) Link

Implemented K-means Clustering to cluster users who visited similar destinations from a training set of 103,000 data points

Designed a framework using Decision Tree to assign any test user to a cluster label and then get recommendations for destinations based on the test user’s Cosine Similarity with other users in the cluster

Technologies: R, R Notebook, recommenderlab, rectools, kknn Parameter tuning and feature selection on Sonar dataset (2016) Link

Optimized parameters of a SVM classifier by hyper parameter tuning using Stratified cross validation and ROC AUC

Achieved a F1-score increase of 9.2% by Recursive Feature Elimination and Predictive Mean Matching Imputation Method

Technology Used: R, R Notebook, ggplot2, glmnet, mice, caret, mlbench, glmnet, Recursive Feature Elimination, ROC Sharmili Nag

Contact this candidate