Dokita - A super fast visual platform for exploring big data
Organisations of all sizes and shapes are drowning in an ever-growing mass of data and there is a lack of tools that make it easy to inspect and explore large datasets that may contain thousands or even millions of data points. Dokita is a visual data platform that maps out the entire landscape of a dataset out allowing organisations to explore large datasets without using complicated databases or overwhelming dashboards. Dokita can securely and perfomantly handle very large datasets in the browser, allowing users to explore and filter data in real-time.
Capabilites
Custom backend and frontend, GPU integration to handle big data in the browser, custom data connectors to handle different data sources
Challenges
Handling very large datasets performantly directly in the browser.
Designing a system that allows users to visually search large datasets via real-time clustering.
Outcomes
A platform that can performantly handle extremely large datasets in the browser.
A fully flexible and scalable system to visually search any dataset.
Searching for data is commonplace in the information age; however, the current paradigm is to search for a keyword and to find an exact match, which is sometimes useful, but when you want to see patterns, clusters and see the relationships in your data, we need a new paradigm to search. Dokita is a data platform for searching visually across your whole dataset with multiple parameters, allowing you to see real-time clusters of data that match all or some of your parameters. Dokita stops you from missing interesting data points and enables you to see your dataset from above, which you can’t do in tools like Excel when navigating row by row.
Dokita has advanced GPU-backed data processing techniques implemented inside the platform to allow you to render and use large datasets in the browser without having to connect to databases or external tools, although we can set up Dokita to use these if you need them. We’ve focused on allowing you to algorithmically navigate your dataset via real-time data clustering and refined UX so you can be intentional but also have full awareness of where each data point sits inside your dataset.
