RESEARCH ASSISTANT | UNIVERSITY OF SOUTH FLORIDA, USA | MAY 2019 – DEC 2019 · Published Regional E-insights report of Tampa Bay region and presented in front of 400 executives and leaders at State of the Region Conference 2020. Whole conference was based on this work. · Scrapped data from Indeed, LinkedIn, Twitter and automated data retrieval from Census Bureau using python and selenium. Used advanced analytics and visualization to characterize, explain and forecast the region’s economy.
DATA ANALYST | NUCLEUS TECH, HYDERABAD, INDIA | OCT 2017 – JULY 2018 · Performed extensive data cleaning and data mining using python to build predictive models and used advanced SQL skills to ensure data is pulled optimally. · Built enterprise-level data integration and data transformation solutions. Created SSIS packages in SQL Server Data Tools with custom event handlers.
APPLICATION DEVELOPMENT ASSOCIATE | ACCENTURE, INDIA | AUG 2016 – SEPT 2017 · Engaged with business groups and partners as part of an analytics team to analyze data discrepancies in ETL pipelines and come up with solutions to mitigate it. · Extensively dealt with Microsoft’s enterprise data and sales-force data forecasting.
Projects
CORPORATE RISK MANAGEMENT: · Collected 45 different corporate crisis instances and around 3.7L tweets regarding those issues and performed topic modeling. Used TFIDF vectorization and an unsupervised ML approach LSA. · Built a model that predicts the time taken to change people perspective on a company that is managing its crisis.
RESTAURANT BUSINESS ANALYTICS - BIG DATA ANALYSIS: · Main objective is to provide valuable insights and analysis to the individuals who are planning to start a new restaurant business. · Used PySpark to read files in HDFS on Databricks platform. Used advanced SQL queries to determine factors that play a role in deciding customer ratings and built predictive models using PySpark to determine restaurants’ success based on the business plan.
WEB APPLICATION DEVELOPMENT USING FLASK FRAMEWORK: · Used CSS styling to the HTML pages and returned these templates using python and flask. · Used a PostgreSQL database to store the data and Heroku to deploy the application online. Data statistics were mailed as an alert using smtplib module.
FEATURE RATING EXTRACTION FROM USER REVIEWS USING PYTHON: · Developed a web scrapper that extracts reviews from an e-commerce website and pre-process it. · Individual product attribute reviews were derived from this data and ratings were assigned based on sentimental analysis to better judge a product based on an attribute. This helps the company to come up with actionable fixes in future products.