SUMMARY

• SAS certified professional with over 6 years’ experience in data analytics and predictive modeling having an aptitude to transform primary data into strategic & actionable insights.
• In depth knowledge of statistical procedures including but not limited to methods such as logistic regression, multivariate analysis, decision trees, artificial neural networks, random forests, time series analysis and forecasting, principal component analysis.
• Conducted coding of qualitative variables and manipulation, complex hypothesis testing and statistical analysis through various statistical methodologies like experimental designs (ANOVA with or without replication, factorial design, ANCOVA etc.), MANOVA, discriminant and factor analysis and statistical inferences in SAS, R and SPSS computing environment.
• Proficient in analytics tools like SAS Enterprise guide, Enterprise Miner, SPSS, JMP, & Weka.
• Comprehensive knowledge of wing - to-wing exploratory analytical support including extraction, data cleansing, pulling data, fitting models, interpreting model outputs, preparing visualization reports and making strategic recommendations to clients.
• Co-ordinated with cross functional teams to prepare technical documentation & analyze the results obtained and also collaborated with key stakeholders to comprehend and analyze the requirements
• Excellent analytical skills and business acumen for understanding the business requirements, business rules, business process and detailed application design
• Strong understanding of SAS coding including but not limited to SAS/BASE, SAS/SQL, SAS/MACRO, SAS/ODS, SAS dataset MERGE’s, FORMAT/INFORMAT, SAS DDE, SAS Functions, SAS TABLES, REPORT.
• Through Knowledge and understanding of Python scripting, Microsoft tools MS Excel( Pivot tables, VLOOKUP, HLOOKUP, FILTERS), MS Access, MS PowerPoint
• Ability to quickly adapt to new applications and platform modules

TECHNICAL SKILLS

Predictive Analytics: Logistic regression (customer attrition), Multiple regression, Multivariate regression, Random Forest, Artificial neural networks, Decision trees (CART, ID3 & C4.5), Ensemble methods, Support Vector Machines, ARIMA, ARIMAX.

Descriptive & Inferential Analytics: Dimensionality reduction- Principal component analysis, Factor Analysis, Cluster analysis (k-means, hierarchical), Discriminant Analysis, Repeated measures, MANOVA, ANOVA (Hypothesis testing), Tukey, T-test (paired, two-sample), Model comparison (Likelihood-ratio), Generalized linear models, PC regression analysis, variable clustering, OLS, Bayesian netwroks.

Dataset Retrieving & Cleansing: SAS(MERGE, ARRAY, UPDATE, SET, SQL, MACROS, LOOPS, SUMMARY, TABULATE, REPORT)

Analytics software: SAS(Base & Advance), JMP, AMPL, SIMUL8, R, Excel, Weka, Minitab, SAS Enterprise Miner, SPSS

Database & Programming: MySQL, XML, Excel VBA, MS Access, SQL Server, SSRS, Python, Basic JAVA, Hadoop(basics)

Certifications: SAS Base Programmer, SAS Statistical Business Analyst - Regression and Modeling Credential, SAS Advanced Programmer

Data Visualization & Presentation: Tableau(dashboards), Weka, SAS-Enterprise, Excel(Pivot tables), PowerPoint, PREZI

PROFESSIONAL EXPERIENCE

Sr. SAS Developer

Confidential, Texas

Responsibilities:

• Analyzed credit reports to determine credit worthiness and performed risk evaluations using logistic regression
• Evaluated and recommended affordable options to mortgagor to help reduce one’s financial loss
• Prepared large datasets by pulling the information from relational database and performed statistical testing
• Utilize advanced multivariate statistical methodologies like multinomial logistic regression, discriminant analysis, principal component, cluster analysis etc., and Bayesian regression techniques to manipulate data and to build predictive models to target desired group of the customers for the company. Used Python and SAS to develop models.
• Performs complex pattern recognition of financial time series data and forecast of returns through the ARMA and ARIMA models, exponential smoothening; Autoregressive Models for multivariate time series; incorporated PPNR modeling techniques.
• Undertook data or record collection, data cleansing and coding, data aggregation and validation, model development and scoring, predictive analysis and postmortem analysis of the model using R, SAS and Excel.
• Analyzed and interpret significant results based on data summaries
• Developed a time series forecasting model and checked for autocorrelation, due to the under-fitting of the linear trend in the time series plot, ARIMA models are fit and analyzed for model adequacy checking, autocorrelation (ACF). Validated the model and forecasted the response variable with one step ahead forecasts and documented the process
• Data is preprocessed (cleansed) to address the class imbalance using SMOTE and spread sample filters in Enterprise miner. Random forest classifier is tuned to maximum efficiency by configuring no. of trees, features, maximum depth, debug and random no. seed and prediction accuracy was analyzed on ROC curves (sensitivity, specificity, precision) & F-measure
• Wrote complex SQL query and views for a relational database(MS Access) to handle ad-hoc business requests
• Used stored procedures to monitor the database for real time updates and tuned SQL queries for better performance

Data Scientist

Confidential, CO

Responsibilities:

• Conducted advance analytical tasks, data mining, predictive modeling, text analytics and clustering techniques to enable customer feedback and improve products and services.
• Dealt with large volumes of data, exploring and identifying data useful for retail business.
• Automated processes by SAS Macros, SAS-SQL to enhance the effectiveness and efficiency in SAS coding and reduced the cycle time of standard analytical deliveries.
• Developed solutions to optimize loyalty/retention campaigns, using customer profiles to design sequential offers to attract customers to higher levels segments
• Utilized logistic regression and other statistical techniques to oversample and clean modeling data sets, derive and reduce modeling variables, transform and develop linear predictors
• Manipulated data by validating, cleansing and/or imputing errors, missing values and outliers, ensuring accuracy, completeness and consistency for analysis and reporting purpose
• Developed reports with ‘Was is’ analysis, ‘What-if’ analysis, Trending analysis etc., executed various models and analyzed results and produced reports in simple and structured format.
• Binary logistic regression model was developed using maximum likelihood to predict probability of customer interest. Model adequacy & Goodness of fit are measured on chi sq.- statistic, concordance, c statistic and final model is validated on validation dataset using ROC and GAIN charts. New observations are scored on the final logistic regression model
• Fitted a linear regression predictive model using six continuous and one categorical variable, analyzing R2pred. Model adequacy is checked using residual analysis (PRESS, studentized, R-student) normal probability plots, partial F-tests. Addressed multicollinearity issues by principal component regression and validated the final model with new observations
• Designed an interactive dashboard application to retrieve data and visual charts from MySQL database (relational database). Dashboard had interactive visualization tools to visually represent the retrieved data (performance metrics) in spreadsheet. Used VBA coding with ODBC connector for retrieving and updating data in MySQL database & documented the process
• Resolved correlation issues in design variables through proactive use of data mining and statistical techniques and used t-test, ANOVA, linear regression (parameter estimation) to understand design failure and monitor key process parameters
• Participated in design reviews with cross functional teams and communicated statistical inferences from regression studies
• Used in-house ERP tools for extracting large-data, preparing reports, data visualization charts(tableau) through SQL query
• Analyzed risk parameters (financial & design) for new products through experimental design, hypothesis tests & validation that drive product and feature enhancements, providing new value to product development
• Created & executed tests scripts for loading database and aided in escalation/ resolution of data issues.
• Constructed a Monte Carlo simulation model for manufacturing plant and decreased the resource utilization by 6%. Analyzed the assembly line performance metrics, resource utilization and addressed the future impediments
• Implemented 2 3 factorial design, built a ANOVA table and checked for model adequacy (normality and constant variance). Factor screening was done using ANOVA, plot of effects, interaction plots. Monitored on R2, Adj-R2, Root MSE on SAS