In this blog post, which is the second one in the series of the “Components for Data Sharing, Value Generation and Intelligence” bundle, we cover briefly the next 2 components of that bundle, namely the Secure Analytics Playground and the Data Explorer.
Secure Analytics Playground
The Secure Analytics Playground (SEAS) is a module inside DataVaults platform that offers the possibility of running different algorithms over the shared data in a secure way.
The SEAS is a module developed for use by the Data Seekers however Data Providers could use it to gain insights into their own data. The module is divided into two subcomponents.
The first subcomponent “Service Analytics Host” is in charge of defining the features of the setup and configuration of the Playground and it also has a UI to manage and define these characteristics.
The user will be able to deploy the playground on a defined server or local, for example on their own laptop. The Data Seeker will have the Playground and Visualization modules in an isolated environment to create their own models or algorithms or to select the ones provided by the platform.
The second subcomponent “Playground & Visualization Host” will allow the Data seeker to execute and track the Machine Learning models. The tracking capability allows the user to decide which one of these will be stored in the Cloud, sent to the Visualization module, or both.
The “Service Analytics Host” will use technologies such as Play for the User Interface, Java for the backend setup and Docker for the deployment. Ansible will allow us to execute the different tasks associated with the deployment on various platforms (e.g. local or on a defined server).
The “Playground & Visualization Host” will be based on Mlflow to execute and track the Machine Learning models, PostgreSQL to storage the results of the models and Apache Superset to visualize the results
Initially part of the larger component called “Query Builder and Data Explorer”, this component is the one which allows Data Seekers to browse into their own Data Space and retrieve, delete or forward for analysis the data assets they have already bought.
The Data Explorer is written in Java, using VueJS2 in the frontend.
In the final blog post of this series, we will be covering the SSE Engine and the Edge Analytics Module