Libraries support a variety of flavors of data ingestion, data prep, automated basic statistics and metadata supporting rapid EDA, creation and evaluation of predictive models, and advanced visualization.
Selected cloud data can be seamlessly transferred to a local (non-cloud) computer, for example a laptop, if appropriate and secure, enabling offline analysis.
The platform is primarily available for Trexin data scientists, but Trexin also offers it to client data scientists who need a powerful turnkey platform for analysis and modeling. We caution clients against using this or any platform without sufficient statistics and analytical knowledge, in order to avoid costly mistakes in evaluation of analysis and model results.
Core Problems Addressed by TDSP
- No easy ways for some client data scientists to quickly analyze their data
- Inefficient analytics knowledge transfer within Trexin (e.g., healthcare)
- Inconsistent analysis across Trexin data analysts, data scientists
- Scattered training resources for more junior would-be analysts and data scientists
Goals for TDSP
- Create an analytics platform supporting collaborative data science, for both Trexin resources and appropriate clients
- Develop and offer proprietary, domain-specific code libraries to accelerate and standardize (as appropriate) the analysis process
- Create a platform that junior resources can use for experimentation and training
