⚡ Rill Developer is a tool that makes it effortless to transform your datasets with SQL and create powerful, opinionated dashboards.
To try out Rill Developer, check out these instructions and let us know over on Discord if you encounter any problems or have ideas about how to improve Rill Developer!
Building out data projects on a local machine has lots of benefits from leveraging local compute with duckDB to having a clear security perimeter around your data and analysis. However, most great analyses should be shared with others to make an impact in your organization. This release brings additional features that are building out a path to an incredible hosted experience with stateful URLs that share the exact view you are looking at with your non-technical (or honestly just busy) coworker.
What does a hosted Rill dashboard feel like? It captures the dashboards you have on your local machine and sends it to a cluster that can handle more data and is available online for others to see and consume. For example, an operational monitoring dashboard that looks at real-time 311 calls in the California Bay area was created locally, pushed to the cloud using project files, and hosted on an accessible URL for everyone. You can check out this hosted example here.
Sharing specific insights with URLs — Our first version of “stateful URLs” in Rill have been built to support sharing specific views across your team. For our first pass at shareable URLs, we focused on making it possible to re-create the dashboard state, or the combination of all of the filters, metrics, time grains etc that make your view of the dashboard unique. Rather than focus on conciseness or human-readability, we will advantage versioning for now. This is because we’re likely to iterate on the dashboard state quite a bit over the next few months and it is important that URLs that you save don’t break.
Histogram improvements refine distribution insights — Our integer histograms were informative, but often not meeting the mark for profiling. People looking at histograms are examining distributions of numerical values in charts and get an accurate and meaningful understanding of the shape of the data. Historically we applied a well-vetted algorithm based on Freedman Diaconis Estimator, but in practice we find that it is bucketing too aggressively for users to see the nuance of their data. A few small changes to (1) the optimal number of bins and (2) where we place the bins have drastically improved the experience and open the door for deeper diagnostics enabled by scrubbing histograms and zooming in.