Solid analyst with many years of experience in market research, segmentation, and statistical modeling. Current areas of focus include market potential studies, potential partner research, and investment opportunities for large institutional clients, pharmaceutical and biotech companies. Ad - hoc reports covering other topics on request.
Databases: Oracle, SQL Server and SQLAzure (including use of SSMS, SSIS, SSAS), PostgreSQL, Access, and the IBM Mainframe.
Big Data Tools/Technologies: Hadoop, Hive, Spark, and some Scala.
Programming Languages: SAS, R, Python;
Reporting Tools (Tableau and SQL): Knowledge of SAS on Window and the Mainframe. Coding in COBOL. JCL, OS/MVS, TSO/ISPF. IMS DB/DC. Familiar with Source Code Management Tools
Data Modeling: Oracle Data Modeler, Visio, Project Management MS Project, SharePoint; Excel. Testing tools for quality assurance. Whole Software Development Life Cycle. Experience with ISO, ITIL, CGxP, and other process standards.
- Performing analysis of market potential for applications of research across different diseases by treatment method. Client would like to uncover breakthrough opportunities for new treatments and find partners to develop startups.
- Working with members of data team, came up with a risk modeling strategy to create expanded risk profiles for various clients in the insurance industry.
- Used the Finite Mixture Modeling technique to stratify patients according to high risk, high utilization, and mortality.
- Developed risk profile segmentation based on medical data enhanced by outside sources.
- Built "Personas" of the various sub-populations of Confidential prospective customers (a combination of pregnant women, their spouses, and families) using Confidential and clustering methods.
- Result provided a comprehensive actionable profile on prospective leads in the client’s email marketing database enhanced by rich 3rd party demographics where matching data was available.
- Working with the CordBlood Registry Sales and Marketing teams, Built model predicting the probability that lead in the marketing database would request more Information on the product. Resulting model provided 4.98% conversion, which is a 100% lift over the baseline rate.
- Top quantile performs at 6.28% conversion rate, which is 126% over the baseline.
- Performed a sentiment analysis of specific brands on Twitter. Uncovered "influencers" for specific brands such as hotels, outdoor sports, etc using graphical modeling techniques.
- Managed a cross-functional team of external consultants, internal IT, and Product Marketing to implement a coupon recommender system for the “ Confidential ” and “Items You Buy” sections of the Just 4 U Coupon Center. Built a recommender system prototype in R, and scaled it to support 7 million customers using Hadoop/Mahout.
- Leveraging social media to engage existing customers and create stronger customer relationships with "influencers" using SAS on Windows.
Lead Data Scientist
- Oversaw the analytics team’s conversion to “Big Data” approaches to data analytics/predictive modeling for Channel partners (Financial Institutions).
- Mentored junior analysts in traditional marketing techniques like segmentation.
- Developed strategies for campaign targeting and optimization of offer allocation for a portfolio of 35 million customers and a history of over a billion credit and debit card transactions using Hive/Hadoop. Implemented logistic regression algorithm for Rare Event Logistic Regression using Pentaho/Weka.
- Developed strategies for campaign targeting and optimization of offer allocation based on Confidential (Recency, Frequency, and Monetary Value). Developed targeting process so offers with the highest expected value are allocated to cardholders with highest Confidential score.
- Also performed lifestyle, and other types of segmentation based on Point of Sale data.
- Provided guidance to Edo regarding best practices for test versus control analytics for promotional campaigns.