Digital Research Platform
The foundation of the Enhanced AI Center is our Digital Research Platform (DRP), which was designed by Research Informatics to make research more agile, efficient, collaborative and secure, while providing higher quality data to researchers. With the DRP, it’s easier to make scientific discoveries.
About the DRP
The DRP is the collective name for our HIPAA-compliant, research cloud infrastructure, which includes a centralized research data lake house, built on Databricks, to enable big data handling with Apache Spark and robust performance with Delta Lake. High-performance computing, including GPU clusters, is available from Microsoft Azure and can be scaled on demand, and machine learning lifecycle management is available from MLFlow. Dedicated research data management staff are available to help ingest new data into the environment, using a customized version of Data Basin to facilitate and speed complex data ingestion into the lake house. Analytics and model building can be performed in Databricks, in Azure or in the user’s on-prem environment.
Benefits
- Data Exploration and Analysis
- Data Processing
- Data Warehousing
- Machine Learning and AI
- Real-time Analytics
- Data Pipelines
- Predictive Analytics
- Collaborative Work
- Interactive Data Exploration
- Fraud Detection
- Scalable Compute
- Optimizing Business Processes
Infrastructure
- Big Data: Most data in the DRP are stored using Apache Spark, a distributed computing platform that can handle very large data efficiently. If you have big datasets that your computer might struggle to load, they can be brought into the DRP for analysis.
- Machine Learning and AI: The DRP uses Databricks, a data science platform that can enable you to use tools such as Pytorch, TensorFlow and MLFlow, without struggling to get them installed properly. They just work.
- Scalable compute: Using the DRP, we can provide the computational power you need for a study, without purchasing it in advance or needing to maintain it permanently. Compute is provided on demand. We can provide a letter of support for your grant describing the infrastructure.
- Privacy: Sensitive data can be stored in the DRP and easily managed in a HIPAA-compliant way. More granular control is available in the DRP then using folders, because access can be controlled down to the level of columns in the data.
- Consistency: Instead of sending datasets to an analyst, we can grant access to your dataset in an environment that allows analysis. Not only does this help to protect privacy, but it also means you can make sure everyone is working off the same version of the data.
- Availability: Your data and analysis are available from wherever you can login, so you don’t have to worry if your computer goes down, or you need to unexpectedly work from another machine.
Accessing the DRP
- Contact us at dataconcierge@kumc.edu
- Access to databases within the DRP may require additional processes and information.
- Once approved, a member of our team will reach out to you with a welcome packet and DRP resources. Learn more by visiting the Artificial Intelligence Learning Hub.