Starting Guide
TeeBI is a performance-oriented API (component library) to enable data gathering, processing and optionally interacting with machine-learning algorithms.
Data can be created manually by code, or it can be automatically imported from many different sources and formats, like databases, files as csv, json, xml, Excel, etc both by code or using editor dialogs.
Data can be optionally persisted to native binary files that provide ultra-fast access, as each column or field is directly stored from a simple array in a single operation. Millions of rows take few instants to load, sort, aggregate or traverse.
Basic summary features like filtering, sorting, expressions, aggregations and multiple group-by consolidations can be done using simple classes and/or visual editor dialogs. No sql is required, but it is supported.
Integrations with the "R" language and Python's frameworks like Scikit-learn provide an easy way to use TeeBI data with statistical and machine-learning algorithms from a common Delphi-style set of classes and components, without requiring to write R or Python scripts.
Data visualization is delivered by TBIGrid, TBIChart and TBIVisualizer controls, both for VCL and Firemonkey, that can automatically display data and summary outputs with no coding required.
A TBIDataset component links to any data object using a simple Data property. TBIDataset can be used like another TDataset both at design-time and run-time.
Plugin classes are provided to choose your preferred 3rd party Grid control (Jvcl, TMS, SMDB, Woll2Woll, etc) and data import components (OmniXML, SuperObject, UniDAC, etc) instead of the default Delphi included ones.
TeeBI can be used in two different modes:
-
Low level (just runtime support, no design-time)
-
High level (design-time, no programming required)
The "Low level" section is an overview of the library base classes.
"High level" describes the editor dialogs to configure and import data without programming.