Browser capabilities have exploded in the past years and with it the possibilities to run large applications without the need of additional software. Still when combining interaction and visualization of larger datasets the number of nodes such as with SVG becomes quite difficult to handle (limited resources). That is why we have been experimenting with solutions to make use of WebGL, mixing rendering and maintaining interaction. In order to be able to render scientific data (as read by geotiff.js) we have developed plotty (https://github.com/santilland/plotty). This tool allows us to quickly colorize and render the data using a shader, making it possible to create interactive and explorable animations. For further data analysis, such as evolution curves through time or larger and complexer plots we have been working on graphly (https://github.com/EOX-A/graphly). In this case for us it was imperative to maintain complex interactivity while using rendered images. This was achieved by combining a second canvas which keeps unique identifier colors for the objects. We would like to present our concepts and experiences of the developed tools.