Candidates About

Data Scientist – Business Analyst - Project Manager

Information Technology Professional specializing in data science & data analytics with solid experience supporting a diverse mixture of business sectors including insurance, investment banking, pharmaceutical, and healthcare.  Expertise as a Data Scientist, SAS Developer, Python Developer, Business Analyst and Project Manager with a record of accomplishment in building and delivering comprehensive IT implementations to production.  Well-versed in all aspects of Project Management, building teams, project planning, managing stakeholders, vendors and customers and providing project updates to the senior team.

Additional Strengths and Competencies

·         Python: Pandas, NumPy, ScitKitLearn, Stats Model, SciPy, PYMC3

·         Amazon Web Services: Elastic Map Reduce—Provisioned Hadoop clusters-ran/terminated jobs, and handled data transfer between EC2 (VM) and S3 (Object Storage)

·         Data Visualization: MatPlotLib, Seaborn, plot.ly, SAS Visual Analytics / Visual Statistics, Tableau

·         Operating Systems: Windows, ISPF/TSO, UNIX/UNIX Shell Scripting (via PuTTY client), MS-DOS, OS

·         SAS:  Base SAS 9.4, Enterprise Miner 14.1, SAS/STAT, SAS Data Integration 4.6, SAS Business Intelligence Suite, SAS Enterprise Guide 6.2, SAS macro facility, SAS OLAP Cube Studio, SAS ODS, SAS Stored Process, SAS Web Report Studio

·         Windows Applications: Visual Basic, MS Office Suite -- Excel, Access, Word, Visio, PowerPoint, Project

·         Relational Database: MS SQL Server Studio, Oracle, Sybase

·         PM Training Course: Developing Complete and Consistent Business Requirements, PMI Registered Course; experience working in an Agile environment

Data Scientist  — General Assembly (2017 to 2017)

·         Data Science Immersive program focused on Statistical Analysis, Machine Learning and other modeling techniques exclusively in Python.

·         Projects include: Predicting Management Liability Insurance Losses for a Securities Class Action Lawsuit; Predicting Real Estate Predicting in Ames, Iowa using Linear Regression models; Predicting Data Science Salaries for an indeed.com Webscrape; Disaster Management by Predicting Survival Rate using Titanic data; Predicting Movie Ratings Using Decision Trees, Random Forest, Extra Trees from Application Programming Interface (API).

·         Topics covered: Logistic/Linear Regression, Hypothesis Testing, Web Scraping using Selenium and Beautiful Soup, Decision Trees, Time-Series, ARIMA, Hierarchical Clustering, Principal Component Analysis, Support Vector Machine, Regular Expression

Key Achievements / Capstone Project:

·         Performed a training data/test data split in order to better manage variance / bias tradeoff in model building process. Performed Lasso and Ridge Regularization on Logistic, Linear Regression Model.

·         Produced Decision Tree, Bagging Decision Tree, Extra Trees and Random Forest to guide the building of a logistic regression model that predicts whether a Securities Class Action lawsuit will settle or be dismissed.  Trees and ensembles determined feature importance to benefit feature reduction.

·         Fit Logistic Regression, Linear Regression Models on training data, using SciKit Learn. 

·         Produced Confusion Matrix and Classification Reporting to visualize the performance of the logistic regression model according to accuracy, precision, recall scores.

·         Using plot.ly produced ROC Curve to visually represent True Positive Rate versus False Positive Rate.  Equally produced visualization of Precision Recall Curve for Area under the Curve.

·         Used non-parametric K-nearest neighbors technique by fitting model on test data and minimizing the sum of squares for the distances between data and finds the corresponding cluster centroids.

·         To address non-linear data, performed DBSCAN (Density-based Spatial Clustering of Applications with Noise) unsupervised learning clustering technique. Visualized the identified clusters from DBSCAN using plot.ly.

Senior Consultant/Data Scientist — Nationwide Management Liability & Specialty Insurance Co. (2010 to 2016)

·         Analyzed the Securities Class Action SCA lawsuit data landscape to ensure the accuracy of the Director and Officer (D&O) predictive pricing and relevant to the underwriting staff.

·         Ran logistic regression models using PROC LOGISTIC to test likelihood of U.S. Corporation would have a SCA given corporate financial data, external risk factors and proprietary and external loss data. Goal was to strengthen prediction power of risk score.

·         Ran PROC GLM to create linear regression, risk score as target variable. Created linear regression models testing dimensionality reduction.  Also feature engineered based on existing data: dummy variables, interaction variables.

·         Performed various T-test analyses to determine whether there is a statistically significant difference among SCA loss data across segments (e.g. subprime / merger and acquisition).

·         Monitored and analyzed Securities Class Action (SCA) related trending economic events to build an accurate modeling dataset used for policy pricing, risk assessment, and actuarial loss triangle development processing.

·         Utilized SAS Visual Analytics to create tile charts, heat maps, graphs, maps and other representations to analyze the SCA landscape.  Uploaded data via VA Laser server.

·         Scheduled SAS jobs via Platform Computing scheduler software.

o    Used SAS Stored Process (SAS Output Delivery tasks) to call SAS Data Integration jobs to compare data and for data transformation. Executed monthly SAS jobs via UNIX Shell scripting.

Key Achievements  

·         Performed text analytics of legal SCA complaints to categorize and segment relevant data points into the predictive model.

o    Performed credit rating analysis to identify the likelihood of public company SCA based on downgrade/upgrade of rating.  Analysis focused on predicting a SCA given changes in Moody’s/S&P credit rating systems.

o    Gathered derivative class action data and default/bankruptcy data for modeling data preparation. 

o    Created an underwriter reference guide that detailed data sources/processing/warehousing/output used for the Predictive Model.  Trained the new underwriting staff in the account submission system that utilizes Predictive Model for Risk Assessment. 

·         Developed a litigation management front-end system that allows underwriting and claims staff to research the SCA universe.  Incorporated Extract Transform and Load (ETL) components using SAS Data Integration in addition to querying/reporting functionalities.

·         Provided analytics and price mechanism that allowed underwriting team to avoid adverse risks at adverse levels and avoided 1%-3% toxic accounts and toxic risks.

·         Streamlined processes that saved the Company over $100,000 in costs, ~10% of a $1.1M budget.

·         Presented at 2015 Insurance Accounting & Systems Association's Educational Conference and Business Show. The topic discussed 'Transforming underwriting with Data and Analytics.’

Additional Experience

Project Manager — Deutsche Bank

Business Analyst / SAS Developer / Project Manager — UBS Wealth Management

Senior Business Analyst — Medco Health Solutions

Business Analyst / SAS Programmer — PricewaterhouseCoopers LLP

Technical Research Assistant — Manpower Demonstration Research Corporation

Data Science Immersive Certification General Assembly

 M.S. Intellectual Property Albany Law School
B.A. Hobart & William Smith Colleges