Welcome to Aberdeen’s project documentation!¶
Goal of the project:¶
The aim of this project is to develop a Python framework which can be applied to any kind of database in order to build an explainable mathematical model using an algorithm based on decision trees to extract the rules which govern the mapping between selected features and a given targeted label. The rules are then presented to the user in form of an image which can be easily interpreted.
This framework is tested on the Aberdeen data. However, as mentioned above, this framework is generic and can be applied to any kind of database as long as it’s available in form of a single or multiple csv-files.
Methodology of data gathering:¶
The Aberdeen Technology Data Cloud (ATDC) research methodology has been honed over 40 years of data collection. Trained research assistants complete interviews each month to ensure that the database is as fresh as possible. The database provides coverage across all industries and company sizes. Aberdeen Group ATDC also leverages inferred content across 20 technology areas ranging from total installed PCs to the likely presence of WAN. This intelligence enables Aberdeen Group to develop scores through statistical modeling and application of data obtained that is then provided within the ATDC. From single-location companies to Fortune 1000 enterprises, we seek the business locations, or “sites”, which have future IT purchases, and the decision makers who have the influence and authority driving the purchase of significant technology installations. Respondents to ATDC’s syndicated interviews are IT professionals who are knowledgeable about the technology present at their locations. In each applicable section, all spend numbers are in local currency.