A patent was awared to Reginald Nosegbe of Valspresso for the classification of stocks based on an innovative theoretical model. To make use of the the classication system Valspresso required a database that contained of all publicly available information for all publicy traded companies. This big data project involved sophisticated database design and an expert system. A software architecture was developed to build a foundation for all future company products.
We developed an expert system that manages the stock portfolio of several clients. The system is based on theoretical framework developed by Reginald Nosegbe of Valspresso. The portfolio includes an advanced trading simulation system that is used to develop and refine trading strategies. The current strategy called Green Zone SelectTM has been yielding returns higher than the S&P 500 at lower risk. Valspresso is now in the process of opening a mutual fund.
The client was looking to develop an electronic version of a paper-based questionnaire, which was about 15 pages long and contained over 3000 questions. The data on this questionnaire had been previously reviewed and scored by hand by a highly paid professional. The results were then dictated on tape and typed by office staff into a standardized report. The scoring rules were well known to the experienced professional, but it was difficult to teach these rules to other professionals. The data previously collected was prone to error due to misunderstanding of instructions or because of intentional deception by those filling out the questionnaire.
A dynamic web-based questionnaire was created to deliver the questions to applicants and receive responses. Since the web page employed dynamic HTML, the questionnaire adapts itself based on the user’s responses. This allows the questionnaire to avoid asking irrelevant questions. It also allows the asking of follow-up questions when appropriate. The result is a questionnaire that enforced a higher level of quality, truthfulness, and thoroughness.
An expert system was then developed to process the responses. This expert system employed a custom built rule-based programming language. This language allowed the developers to more quickly integrate the system’s components and allowed the client to define the rules in the way that made the most sense to them.
Developed for Johnson Controls, H2Optimizer is a system that performs sophisticated modeling of a municipality's water usage bills then makes recommendations on the replacement of water meters. The installation of more accurate water meters yield higher revenues for the municipality. Because the installation of new meters is an expensive undertaking, H2Optimizer performs the financial calculations necessary to demonstrate the return on investment. It also enables the engineer to customize the solution to yield the greatest benefit at the lowest possible cost.
A project management solution was developed for Johnson Controls that enabled hundreds of engineering teams to manage thousands of building efficiency projects. A scalable, enterprise architecture was implemented to support the growing suite of tools. The system included about 10 different modules including budgets, costs, project plan, team management, security management, document management, and collaboration. The system receives data from several different databases within the organization. The system has become a central hub of data for new projects. REST web services provide new tools easy access to this centralized database.
Another project for Johnson Controls is for executives who need a rolled up view of current and past projects. A business dashboard provides ask to on-demand reports. Executives can choose one or two dimensions for analysis and one or more facts (data points) for reporting. Reports can be customized, saved, and shared among users. This system is part of their routine performance reviews.
Harrell's is a nationwide producer and distributor of customized fertilizer blends. Their sales representatives are very mobile and require the ability to place orders for customers while on the go. CogniVista developed a custom mobile application that allows sales reps to view pricing information, place orders, check orders, and view reports.
Created a prototype of a new federal standard for data analysis of UI wage records. For the prototype 172 million records were analyzed to produce a 120 GB SQL Server database. Microsoft Analysis Services (OLAP) was used to transform the database from relational to multidimensional. An ASP.Net web application reads multidimensional data to produce charts, maps, and tables. A sophisticated custom query engine was created to allow the user to query the OLAP data.
Created systems architecture and prototype of a peer-to-peer medical record research and e-commerce system. Led team of developers in incremental implementation of the system. Built a service-oriented workflow system, a peer-to-peer architecture, and an e-commerce website. A research network was then implemented on top of those base services to integrate all the systems. This system was based on a patent held by Thomas Leiper of Shepherd Medical Solutions.
NavDog.com map search represents an evolutionary step in design concept and user interface that combines the advances in modern Business Intelligence with consumer information need. NavDog uses geography as the common platform that connects people with goods and services, ties market silos together, and exposes the natural contextual relationships that exist in our world.
NavDog proprietary technology in the mapping layer creates a seamless real-time experience. When a user interacts with the map, the system waits for them to finish their action, queries the NavDog database for content changes and brings them to the map. NavDog shifts the paradigm from a search for specific information to an information browsing experience with the ability to engage the user in virtual exploration of real places. Unlike current mapping systems that require a new topical search for each new subject, NavDog uses machine learning to derive the intent of the user.