We implemented unsupervised exploratory factor analysis (EFA) to help explain risk tolerance in consumer investing for an online assessment tool. This resulted in an online assessment tool that is currently being used to segment consumer risk tolerance for different types of investments. Our analysis was performed to build both a European and US version for segmenting investment risk tolerance.
Examined segmentation and prediction in advertising spend by conducting both unsupervised EFA and supervised structural equation modeling (SEM) for a Fortune 500 multinational medical devices, pharmaceutical and consumer packaged goods manufacturer (> $70 billion yearly revenue). This project used advanced dimensionality reduction techniques (EFA) and prediction techniques (SEM) to discriminate target audiences and predict their behavior related to advertising spend. Our findings showed that the company’s current marketing campaign spend was not related to or resulted in increased customer loyalty. Thus the company needed to rethink their marketing campaign spend as related to increasing customer loyalty.
In collaboration with our client who works with a major U.S. retailer (> $40 billion yearly revenue), we produced a methodology and statistical model to accurately estimate real-time customer counts and traffic patterns. Predictors were used to segment customers and predict in-store purchase patterns. This procedure has since been modified to generalize across other avenues using their technology. Due to this success, we continue to work with this client to improve the validity of their data and statistical modeling.
Conducted advanced psychometrics and survey development using EFA, SEM and FA for a large non-profit organization (> $3 billion yearly revenue) to survey and segment what services their customers would find useful, their perceptions of this non-profit organization, and what services the non-profit should work to improve. We also created geomaps the non-profit used to segment their customers so they could be better served. For example, combining customer needs data with a list of the non-profit locations allowed this client to better determine what services needed to be added to each location.
Our data science team used a novel approach to forecast economic depreciation of car equipment for a major U.S. automaker (> $80 billion yearly revenue). In this case care was the “car” was the target audience segmented by care equipment. Prior to our work, a data-driven modeling approach had been unsuccessful due to the difficulty in estimating car equipment value. Our model overcame previous limitations by allowing the auto-maker to quickly and accurately estimate car equipment values, which proved to be financially lucrative for our client. Due to this success, we have since developed numerous other statistical forecasting models for this client to improve and automate the modeling process.
For Organic Research Corporation Developed a PHP web app called Physician Image Labeler (PIL) Software, which provides tools for managing a hierarchy of personnel performing image annotation and labeling input for medical imaging machine learning processes, complete with reporting and data export features.
We implemented unsupervised exploratory factor analysis (EFA) to help explain risk tolerance in consumer investing for an online assessment tool. This resulted in an online assessment tool that is currently being used to segment consumer risk tolerance for different types of investments. Our analysis was performed to build both a European and US version for segmenting investment risk tolerance.
We built a custom e-commerce integration work implemented in PHP as extensions to the Magento platform.
Built an iPad survey app for Altria (owner of Philip Morris) to help with data gathering at trade show for the convenience store industry. Booth staff could point an iPad mini at guests when greeting and scan name badge and load all of their personal contact information to populate survey. Complexities of this project included tracking a guest’s participation across seven separate surveys for different product lines spread throughout the show floor following a required entry survey. Collected data is stored locally and published to a remote repository via a JSON web service interface when Internet connectivity is available. Also developed reports providing structured data export for the collected data.
We designed and implemented database backend for Marcus Theatres hotline phone service used by customers to find times and locations of movie screenings. Created automated service to load and parse data provided by area theatres containing movie titles, venues, screens, and times. Provided SQL stored procedure interface to Interactive Intelligence I3 IVR system to query and present information in real time to callers.
We developed an ACP program for Stone Wheel Wholesale Auto Parts to assist call center and in-store representatives with order taking and customer service inquiries. Windows desktop program integrates with Interactive Intelligence I3 CIC system to obtain caller ID information for incoming calls and query for account information with a database that cross-indexes known accounts and contact information. Database of customer contacts is synchronized with an IBM DB2 database in an AS/400 environment using a custom service I developed. App provides functionality to update and append database in real time, and includes integration with legacy text terminal application for quick order entry and retrieval.
Upgraded and updated an app built in Microsoft Access and Visual Basic for Nestle Foods to track their marketing divisions’ food promotions. This included major work with updating data processing routines in Visual Basic as well as integration with a remote Microsoft SQL Server database.
Took over development of a small custom health and safety reporting web application for Johnson-Diversey written in ColdFusion and corrected numerous flaws, filled in missing features. Later served as lead developer to build and integrate full safety audit application and custom Key Performance Indicator (KPI) system for Diversey. By the time the company was acquired by Sealed Air in 2011, we had fully rewritten all three integrated applications for the Microsoft .NET platform and greatly expanded the features to include various custom report tools, rich user management, automated tasks, and graphical dashboards.
Lead architect and developer for APMGI Branded Products (BrandedProducts.com) family of e-commerce businesses from 2007 – 2010. Built a comprehensive B2B and direct to consumer platform for premium branded merchandise. Included custom intranet and extranet integration and dealer reseller web portals for Jack Daniel’s and Harley-Davidson.
Implemented fixes and new features for searchable knowledge-base content management systems. Lead design and development efforts of custom web, desktop, and mobile application development, as well as configuring and maintaining onsite managed hosting services. Performed custom integration work on many popular platforms with focus on Microsoft .NET, XML web services, and SQL.
Lead developer for Midland Health Moxie health and wellness screening application used since 2011 to coordinate client screening events and input and retrieve participant biometrics via PC and mobile devices. This greatly reduced errors in data entry and processing and improved operational efficiency. Included custom integration work with Alere’s SAML SSO framework and custom reports via Crystal Reports via an extranet portal. Also included custom two-way encrypted secure data interchange with Unity. Currently updating this application in Python/Django.
Designed and built integrated web applications for PMCA (Pennsylvania Manufacturing Confectioners’ Association), AACT (American Association of Candy Technologists), and The Manufacturing Confectioner utilizing the dtSearch content platform and Miva Merchant ecommerce platform to provide searchable content and purchasable digital downloads, as well as online subscriptions, membership, and event registration services.