Data analysis is inherently multi-step. Often, teams must transfer data across databases—sometimes preserving its structure, other times transforming it along the way. These sources may reside in the cloud or on-premise, adding further complexity. To streamline such operations, an intuitive and efficient interface is essential.
Furthermore, data science professionals typically export processed datasets as .csv files, including test data, which is often shared in the same manner. These files are saved locally or on virtual machines. They may also reside in cloud storage environments linked to serverless infrastructure where tools like Jupyter are remotely accessed. In addition to leveraging large public datasets available online, practitioners can register and share their own public or private datasets using built-in libraries provided by TensorFlow and PyTorch. These libraries however impose a steep learning curve due to framework-specific boilerplate apart from rigid formatting requirements making it complex for beginners.

With Cliosight, users can orchestrate a sequence of data operations—either serially or in parallel—to achieve targeted data quality and analytical outcomes. For example, data can be ingested into an embedded database for real-time visualization on dashboards, or routed through external processing pipelines for validation and enrichment prior to migration into a cloud-native datastore.
Reports in Cliosight can be used for sharing datasets. They can be updated in an external application's code to be written back as a new report or as additional rows in an existing report. The major advantages of this approach is that, users can conviniently share multimodal datasets in a collaborative work environment with diverse technical skills. Also, by applying role-based access control, actions on that data can be restricted by the resource owners.