SUMMARY:

• More than 7 years of experience in Data Science and Big Data Analytics. Passionate about Text Mining, Social Media & number driven predictive analytics.
• Senior Business Analyst with rich experience in progressive analytical roles in different sectors; proven leadership skills.
• Highly proficient in Python, SAS, R, SQL, Microsoft Office Suite; superior problem - solving and communication skills, highly organized & flexible.
• Track record of understanding and delivering project requirements, managing expectations, using business frameworks, communicating with clients, leading cross-functional teams and meeting sensitive deadlines.
• Experience in developing statistical models for Segmentation Analysis for online scenario
• Experienced in Segmentation techniques - CART(Classification and Regression Trees), CHAID(Chi-squared Automatic Interaction Detector), Decision tress(Random Forest and GBM) and K-Means clustering
• Involved in data analysis as a requisite of various ad-hocs & dashboards encompassing various transactions data metrics rendering effective solutions to the client
• Data presentation using MS Office & Tableau(at times) to the Client helping them make data-driven decisions effectively
• Have high level of experience in predictive modeling along with strong analytical, numerical skills and a creative flair and enthusiasm for new ideas and concepts.
• Regular attendee at various national and international analytics conferences, hack-nights and an open source enthusiast
• Good at validating hypothesis, building statistical models, text mining, Natural language Processing (NLP) and Lead generation from Social media sources like Facebook, Twitter etc.
• Actively involved in data extraction from one of the largest databases(Teradata) using an optimized SQL code
• Experience in databases such as Oracle 10g/9i/8g, SQL SERVER 2000/2005, DB2 and Teradata
• Experience in SQL Queries, Creating Views, PL/SQL Stored Procedures, Functions, Triggers, Cursors as per the business needs requirements
• Ability to meet deadlines and handle multiple tasks, decisive with strong leadership qualities, flexible in work schedules
• Excellent interpersonal and communication skills, technically competent and result-oriented with strong problem solving skills
• Good at analyzing collected data to validate and improve machine learning algorithms. Can generate almost real time analytics with scripts in R and Linux shell scripting
• Carried out end-to-end project shaping experience (developing data, building predictive/ explanatory models, develop strategies, monitor/ validate strategies)
• Design and build analytics products using both structured and unstructured data
• Monitored social media conversations based on business objectives and providing high quality actionable insights to clients to make informed decisions

KNOWLEDGE PURVIEW:

• Consumer behavior analytics to identify response, attrition and uplift
• Marketing mix analytics to identify effectiveness of multiple channels of marketing and allocate market expenditure optimally
• Sales force analytics to provide insights on issues like optimum territory size, optimum product bag size, monthly target forecasts
• Web analytics to understand user & competitor behavior on the web and consequently generate more leads and sales
• Credit risk model to identify cycle defaulters
• Credit/Debit card fraud detection models
• Revenue model to identify long term profitable customers
• Churn models to identify customers likely to discontinue the service
• Collection models to plan collection strategies for defaulters
• Optimization models to automatically allocate the telecom routers
• Market basket analysis for product segmentation
• Customer segmentation using customer historical behavior
• Identify the key performers
• Attrition model
• Estimate property value based on neighborhood properties
• Implementing Linear regression, Logistic regression, Decision trees like Classification and Regression Trees (CART), Support Vector Machine(SVM), Random Forests(RF), K-Means clustering, Local Regression, Robust regression, Panel Regression models, Text Mining/NLP, digital signal processing, Bayesian modeling, projection on latent structures, SVD, etc.

TECHNICAL SKILLS:

SAS: (9.3/9.4) (SAS BASE, SAS Enterprise Guide, SAS Enterprise Miner, Analysis Studio, Advanced SAS)R (Used various packages in R for statistical analysis and modeling. Also used it for creating dynamic visualizations). Integrated R with MySQL database for generating real time analytics

Microsoft Excel: Extensively used it for data crunching and basic analysis tasks like creating filters, pivot tables etc., Also created macros for data automation tasks

Statistical Analysis Tools: Reporting, Visualization, Tableau

Databases: Oracle 10g/9i/8.x, SQL Server 2000/2005, MySQL, MS Access

Modeling: SAS, R, Python, Weka, RapidMiner

Languages: PYTHON, R, SQL, PL/SQL, VB, HTML, XML

OS: Windows 95/98/2000/XP/7/8, DOS, UNIX, UBUNTU 12.04, CENTOS

PROFESSIONAL EXPERIENCE:

Confidential, Richmond, VA

Responsibilities:

• Supervision and leading a team of three analysts to accomplish the project.
• Conversion of unstructured email threads between the client and customer into structured format. Separation of customer and client emails. Likewise the chat between client and customer was also segregated.
• Data cleaning and Meta data generation through Python. Whereas the data structured data crunching and statistical analysis is done using R
• Generating sentiment for the customer chats and emails. There by ranking customer care executives based on their performance.
• Categorization of the chats and emails into various categories, thereby identification of major reasons for client dissatisfaction. Employed supervised classification, LDA (Latent Dirichlet Allocation) and TF-IDF for the same.

Tools: Python, NLTK, R, Shell Scripting and Microsoft Excel

Confidential, Stamford, CT

Responsibilities:

• Building the entire recommendation systems code from scratch
• Trying and testing various formats of recommendation systems (Collaborative Filtering, Content based & Hybrid) based on time and memory constraints involved. A variant of recommendation system should automatically kick-in based on requirement
• Research papers on recommendation systems built for Netflix were thoroughly read and understood to in corporate them into the code
• Entire code is developed in Python using Numpy extensively. Other packages were avoided as performance was of primary concern.
• Using Shell scripting to structure data into the needed format. SAS is used to test the performance of Recommendation systems and compare it with other traditional statistical models.
• Received client appreciation for implementing recommendation systems in short time span with good accuracy levels

Tools: Python, Numpy, SAS, Shell Scripting

Confidential, New York, NY

Social Media Analyst

Responsibilities:

• More than a million tweets are generated on every Confidential match day. Since social data comes at a cost after a certain point, did keyword engineering to identify important keywords for data pull
• Data pull is carried out only for above identified keywords to save data costs. Likewise keywords are split into green and red keywords based on the percentage of leads they provide
• Identified spam tweets. Built an automated system which would identify spam tweets. Like data cleaning process has been setup where words are reduced to their root form, removing special characters etc.
• Hypotheses around identifying leads have been formulated and tested. Finally came up with a scoring mechanism which produced decent results
• Identified sentiment of the tweets, intent of the tweet based on supervised classification techniques. Above data points proved to be very useful in scoring tweets.
• Analyzed more than a billion tweets to identify the difference between consumer perception and brand’s marketing efforts. Also analyzed U.S historical demographic, user description data to identify the potential leads.
• Identified reasons and proposed solutions for the high churn rate of super-engaged Latin American customers. Generated and analyzed regular reports on client’s KPIs and provided insights
• Automated the visualizations for the entire data in R with ggplot2. The data is pulled from a noSQL (Mongo) database and stored in MySQL database for connecting with R.

Tools: Python, NLTK

Confidential, New York, NY

Statistical Modeler

Responsibilities:

• Interacted with the client along with pre-sales on understanding their pain points and business problem
• Executed Univariate and Bivarite analysis, reports generation, visualizations etc.
• Understood the data and was part of feature engineering and data cleaning process like outlier detection.
• Tried various classification techniques like logistic regression, SVM (Support Vector Machines) etc. Likewise tried decision tress(Random Forests and GBM- Gradient Boosting Machines) for more accurate results
• Built classification model and delivered to the client satisfaction. Analyzed data from client to validate and improve the existing algorithms
• Already in place at clients end.

Tools: SAS 9.3, Microsoft SQL Server, R and Microsoft Excel

Confidential

Market Research Analyst

Responsibilities:

• Conduct comprehensive primary and secondary research, followed by analysis and interpretation
• Preparing customized research deliverables to address the requirement of clients as and when required
• Preparing statistical database such as end-use industry segment growth rates, GDP growth rates etc. for studying and analyzing market dynamics thereby forecasting on the basis of same.