Bokeh is an interactive visualization library that targets modern web browsers for presentation. Its goal is to provide elegant, concise construction of versatile graphics, and to extend this capability with high-performance interactivity over very large or streaming datasets. Bokeh can help anyone who would like to quickly and easily create interactive plots, dashboards, and data applications.
The development pace of Bokeh is contingent on external funding opportunities, so this roadmap does not mean to imply any specific dates or deadlines. Listed below are the major areas of work the Core Team considers important. These are tasks and projects that we would like to do ourselves, when time and resources permit, or that we would be happy to have done by new contributors.
Bokeh currently can use WebGL to accelerate plotting for larger datasets in some limited situations. In addition to not fully supporting some use-cases, the internal cost is having to maintain two separate code paths. It would be desirable to convert Bokeh’s rendering to always use WebGL whenever possible, so that better performance for larger data sets is always available, and to streamline maintenance for the future.
By and large none of the core contributors have been JS or front-end experts, and for some time the client side BokehJS library was an implementation detail. There is more interest to use BokehJS directly, and with attention from experienced front-end JS developers, BokehJS could integrate better with existing common web frameworks and tools than it currently does. This includes making much more accessible documentation for BokehJS itself. These improvements would also help increase the potential pool of core developers interested in maintaining BokehJS over the long run.
Bokeh currently has some minimal theming ability, that can specify properties for types. It would be useful to have more targets and sophisticated ways to theme individual objects, or collections of objects. Additionally some work is needed to improve or create visual assets that work with different themes, and to make DOM elements easily themeable as well.
Bokeh has never had the benefit of any dedicated web or visual designers, and this is one area where it suffers compared to some other projects. Visually, it looks much less “designed by engineers” than it used to, there is still much room for improvement. Another idea, if Better Theming Support is accomplished is to have visual designers help create a set of attractive themes that can be immediately available to users.
Some users have expressed interest in the ability to make different views of a plot transform smoothly from one state to another. (Compare to “tween” animations in GGplot). This would afford the ability to create visually attractive “data stories” by scripting smooth transitions that connect one view of data to another.
The Bokeh documentation has grown quite large and has much information, but it seems to be the case that many users still have difficulty finding the appropriate documentation or examples, even when they exist. The project would benefit greatly from experienced help to organize the information architecture of the docs, point out missing areas, etc.
Bokeh can display mathematical symbols and formulas by creating and using custom extensions. Users have expressed a desire for this capability to be more directly integrated, e.g to draw math text for axis or tick labels, without needing to resort to a custom extension.
Bokeh supports various sorts of hit testing for selection and hover on a per-glyph basis. It would be good to make support more uniform by adding the requisite hit testing wherever it is currently missing. Additionally Bokeh needs a policy-based mechanism for dealing with “multiple hits” i.e. if the mouse hovers over 50 points, a way to show just one, or just the top N, or even some aggregation or summary as the hover.
It would be very useful for users to be able to synchronize and manipulate arbitrary data / variables across the Bokeh/BokehJS boundary. Currently only DataSource columns and defined properties are available.