Skip to main content

Extended Statistical Information System

Exploitable results

The industry is constantly searching to discover new ways that would improve and offer user-friendly data analysis, in order to arrive at useful conclusions from collected data. This need has resulted from the plethora of collected data that exists, handled usually by non-experts who are not familiar with data analysis methodologies and the interpretation of their results. In addition, there are many data providers who sell their data and are interested in a standardised and user-friendly way of data extraction and analysis, which can be given to their clients and in this way support their marketing strategies. Thus, under this project's framework, an extended statistical information system (X-StatIS) will be developed, which will: - Assist and/or guide the non-expert users, through an interactive user-friendly interface, in order to select and apply the most appropriate data analysis method. - Help them proceed with the interpretation of the generated results, by providing the users with sufficient information and guidance in order to evaluate and understand the outcome of the analysis. - Be flexible and allow smart adaptation of software usage on several statistical databases held by data providers. - Be able to work as regular statistical software with any data set not restricted to specific databases. - Have a standard interface thus allowing third parties to extend its functionality by developing new data analysis methods capable of plugging into the proposed system. - Be based on the most modern and advanced techniques and make use of state of the art IT methods. - Offer a European oriented solution and address a market area, which is currently dominated by overseas companies. The project is targeted at covering the needs of data providers selling data, as well as the needs of end-users dealing with data analysis. This extended statistical information system will consist of the following components: - Main module having a standard application programming interface (API) allowing integration of statistical analysis modules developed under this project or by third party developers. - Statistical Advisor Module. An open and parameterised module giving the feel of a semi-expert system since it will interact with the user and other system components in such way that it will hide the complexity of the statistical expert knowledge required for data analysis. It will guide the user "smoothly" and intelligently to the appropriate method without the need of the user being a statistical expert, in order to understand, and apply the methods offered by the system and interpret their results. - Guidance Tools to let expert data analysts create scenarios to be followed by the Statistical Advisor Module and by non-experts. - A library of statistical methods. All the statistical methods will use a standard API allowing classification of the output in order to be used as input to other methods. - Scripting module for automated analysis of repetitive situations, integrating macro instructions for data choice, methods applied and presentation of results into a file, which can be executed without user interaction. - Analysis maps showing the overall structure of data analysis sessions. Analysis maps could be considered as a history of the analysis, and it could be used to return to previous data analysis steps. - Knowledge database. It will contain the parameters and rules associated with several statistical analysis methods. These will be used by the statistical advisor module for guidance in the selection of data, methods, and manipulation of results. - Metadata wizard. This module will guide the data providers in the description of the knowledge database parameters characterising the data set variables. The advantage of this module is that data sets from any source could be used by the system, making it capable of suggesting the appropriate statistical method for analysis.