Jeffrey Heer

There are potentially related issues with different gradient rendering (at least in Chrome) between SVG and Canvas. I do not (yet) know how to parameterize the rendering to get pixel...

I'm not referring to the different geometries discussed in #2365. Rather the color ramping / blending can differ between renderers, assuming a fixed geometry.

Flagging label transform author @chanwutk to see if he can take a look at this.

Right now the locale settings do not support time multi format specification. However, you can do this by setting the `format` property on an axis. See https://vega.github.io/vega/docs/types/#TimeMultiFormat Here's an updated...

Yes you could! Here is an [editor example](https://vega.github.io/editor/#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-TGg1mKVSiuHI9Hg3GEzENS6dXrDcbZDWs55qN4oUULgWA491lUxk8XlAEL4tYTiVn-jFSh2oOWAT5HMh4UvIKlWDJEzh0axkJmIoxQVpD8eSZAAF5LLTAbPd3NFBZJVIZYuktTcDCOB+QDCMRAiCYKxow8b7lAyaCsKKSQF+vwYk4AKqBoiCFn6qKljMcxFOSW4rjEa4yBuGbXjue4XskR4nn8Z47JedF3qkAFdj28zxO+KwIXRP5-nIOCPkBRQgee8jgZBDaas2KStrK7aIdkyEHHmyCmOYYzYde86iYwd6EaGxHrpudGUVBTE6QxWjYr48TMfeB40VeWb8eg-5CYB+QfBO3DAuJkn1kUjrOq6eDurxFGYqpnyiPIADMABsACMI4PGOIDWOgpTTp8WQ4YCcIuguW7UMk6jUbR15uR5YC2X4LmkniEHVkU7DcAQv6EWeADWADCfmtWAbjXuFeDFUUMGpnBDJafl16ThNUAEXRZUVU5VWub+7mCXVdmNb8zURlGRQutiUZ0WNLoAAp2A4TiuJdzBLZAjZarqclGgpcijhAzyZUYLoHHu-h5bOpKeAZekGU940AiF8ZhRFdG4vi+TAVuEbwNwag7DhRiwOSqDtX4D4BkYmxvHj5PZLA6LcD1yAAOrqJonkUL4R4WEpMaQNwtjwMwBHs5zOwDC4PPBnTqSMwBHM+FzRQAMSiIlAAc2jsMMkudog6AVoGkDy4rUCCyCxVDjGYCW+sNthHblv-dYGDJD1oNgF8kw6bMAIu27hkG4CRMk1u2KpASnkib5-nRJLkBYzj1O85T2xk1bUCLp5i0AouEmQHWUGQFNabwTr2TC4b2cxCtgWF29sn6l9JpyGXBP6XjhVTsB7cKLXMQIy6GLhYpJ6-LAesG4+xsAmb8AW6PtP04zLPoXLovcwvRj8+ggsUmvCti3QEubxOS8d9PMQq+rmvayf+e0r2hsX0U6KYg08LHQhg4kg7TwgI8QA), with a 0.24 threshold chosen so that one axis adjusts to zero and the other doesn't. Vega-Lite could generate code like this...

Yes, I plan to merge the monorepo to master sometime in the next week. Everything is working, just want to double check everything and update the documentation / README.md files.

As you've noted, this is not something that either Vega or Vega-Lite currently supports. We'd be happy to consider a PR along these lines, but unfortunately I don't expect this...

If anyone in the community is interested in working on this, the underlying regression methods are in the vega-statistics package. See https://github.com/vega/vega/tree/master/packages/vega-statistics/src/regression If we have standard error calculations supported, then...

Thanks for checking in. No update yet, but I plan to look more closely very soon!

I took a look, and the issue on d3-color (d3/d3-color#97) has not yet been resolved. Once that is fixed, we should be able to update Vega as well.