Featured

Vibes: more dynamic theming..

Yeah, I'm full of teasers lately. But hey, presentation is a big part of people consuming your data content. Having more levers to pull (even if you never use them) is never a bad thing IMHO. Build a bunch of stuff & forget "what depends on what"? Viewer or Builder: take

Featured

Data-driven theme params

Formatting nerds unite. 👀 I'm a dashboard formatting nerd, so being able to parametrize *everything* is great. Diff filter values get diff color schemes? Yes. Change the viz? Change the tab? Yes. Yes. Spent a couple hours building this test board & am pleased with how putting stuff together "feels". pic.twitter.

Featured

Storyboards & custom views? Nah. Snapshots!

Take a snapshot of the state of all parameters, sizes, locations - and then swap between them like dashboard tabs. When you app runs on "everything is automatically a parameter", this can get very interesting. Storyboards & custom views? Nah. Snapshots! Content untouched but all params, filters, coords, sizes, selected tabs,

Viz *inside* table grids? Yes.

That about explains it. ;) Dragging a compatible already-made viz on to a grid and watch it "weave by dimension X"... Experimenting a bit. Taking an *existing* viz block & "weaving it" across a compatible dimension *inside* a separate data table (regardless of data source)... Hmmm... 👀 pic.twitter.com/nZiD1GM8aR — Ryan Robitaille

Tweets Featured

"Code tables" as general purpose escape hatches...

Write some backend Clojure code - the system interprets it as a SQL table, use it as a SQL table. Filter, join, aggregate, etc. Better yet, save the "code table" as a view -and in production your end-users can use it and never even know it's running code. Disclaimer: This

Tweets Featured

"Visual History" for views and queries

Also works as a great place to put (100% optional) GPT recommendations, if that is something you're into. (but like I said, not a prerequisite - even the base viz recommendation system is 100% not AI enabled and just locally calculated on your machine) Just a quick dev-log consisting of

Drill Down / Drill "Through"

Quick Twitter video of it in action. Things are getting interesting! Drill-down / drill "through". Any aggregate can be reduced to it's base group-bys, even hopping down though nested sub-queries if need be. The spawning summary grid becomes a "super filter" for the drilldown grid. Filter on a row, a single

"Pivot-by/this" 🐇

Explicit pivoting should be a first class feature in all SQL tools, right up there with group-by. Check out these quick video demos below. 0:00 /0:55 1× More visual SQL tools need explicit "pivot-by-this" functionality. Pivot is always dope. Bonus points: Pivoting by *another SQL query* - works

Tweets Featured

Still Cooking - New viz reco flow & "easy" field swap for recos

It's been a minute - good news though, I am now working on this project full-time. Excited to get it in people's SQL loving hands. But to tide you over before some more substantial posts, here is a couple of really cool things I have been working on (plus some

Tweets Featured

Teaser - "Quick Filter" Shortcut

Filtering a table or query-set by one of it's resulting fields is super common and while I have a (wonderful) system in place for this - I wanted to make an alternate "quick(er)" method of doing it. Here is a GIF of the "manual way" (which is still damn

Tweets Featured

Teaser - Column re-ordering and re-sizing

I don't know about you, but I can't stand data apps where the result-set allows re-sizing by dragging the edges of the columns over. I know it seems intuitive, but too often it just results in frustration - either due to the drag target being too small, too hard to

Tweets Featured

Teaser - Sub-query "Lineage Lines"

In a tool that's all about creating sub-aggregations & re-aggregations (or slicing and dicing) of tables - it can be pretty important to understand how all the queries are related (if at all). While not something you'd have on in a production dashboard - I added some toggleable "lineage lines" to

Tweets Featured

Teaser - Testing how themes are working...

"User Params as Theme System" is actually working out well, I think people will enjoy being able to pull themes out of previously saved dashboards and swapping them around. ;) Check out the very short Twitter video at the bottom to see how they are applied... * Browse the existing dashboards * Find

Tweets Featured

Dev Update: Scrubbable User Params = Themes

Quick update on a "Theme System" - I decided that it made more sense to have a scrubbable user parameter system and use that for theme parts instead. Not everyone likes my color palettes so its a win/win. "Scrubbable User Param System w/ Theme Hooks" > "A Theme System" (damn,

Tweets

Teaser - Visual Browsing of Recommendations

User-configured recos can now be browsed in a much more fluid (READ: way less pain in the ass) way. It feels good and it reasonably fast (it is making lots of DB calls to preview with real data, after all). Visual "recommendation" browsing in CanvasQL, my "canvas based" dashboarding / data

Tweets Featured

Teaser - Drag and Drop CSV Import and JOINS

2 short videos where I import 2 CSV files, join them together - and create a simple map from the resulting table... ...and here is using the viz recommendations to draw a US county map based on the joined table. 2/2 pic.twitter.com/qYiqqWCU4n — Ryan Robitaille (@ryrobes) January

Videos

Pre-Alpha Walkthrough

Had some hiccups (trying to render 300k points in vega-lite, oops) - but other than that this is a pretty nice little walk through of basic features and concepts...

Tweets Featured

Teaser - Conditional Formatting Dynamic Color Heatmaps

Added a pseudo "value" to be used in conditional formatting - select any colorbrewer scale and direction - or use the default "temperature" scale. It'll even respect conditional formatting logic - meaning if you don't want it to apply to all the rows - like, let's say you only want

Tweets Featured

Teaser - "Compound" Block Suggestions

Quick video of how the suggestion system allows for pretty much arbitrary block views to be recommended and allows the users query data to be plugged in. It feels almost magical to be able to quickly place simple (and not so simple) viz on the canvas while browsing. Like I'm

Tweets Featured

Teaser - Conditional Formatting (via SQL chunk logic)

Something else I always wanted in a tool like this - easy (and flexible) conditional formatting (full row or specific fields) for my grids. Drag a selected value into a "highlight target" to create a quick (SQL based) rule and then customize it as needed (via code / scrubbers). pic.twitter.

Tweets Featured

Teaser - Drag in "Viz Shapes"

Define potential "Viz Shapes" in the backend config files - those will be compared to the live queries that you are slinging back and forth on the canvas. Matching ones will be draggable as pre-configured blocks that run downstream from the source query block. Not designed to be perfect -

Tweets Featured

Teaser - SQL sub-block lineage

More late-night canvas SQL fun. Not just UI candy. In a system about reactive SQL sub-queries - the "lineage" of a query/block and how it relates to the other blocks is huge. Upstream/downstream, how filters cascade, etc. A neat take on the old pivot table paradigm IMHO. pic.

Tweets Featured

Teaser - Viz Layer Combining

Viz layers. Building things interactively using the canvas as a workspace surface to separate / examine / combine items. Instead of "query mode, then viz mode, then filter mode, then dashboard mode"... Why can't we do all at the same time?#sql #dataviz #clojure #nomodes pic.twitter.com/WJA41mYTdP — Ryan Robitaille (@ryrobes)

Featured

Data Communication is all about unpacking and repacking

A concept I can't seem shake off. I posted a ~2min teaser video to twitter last night about a new feature wherein: * I created a query peeking at an entire table (a select-*) to see what we had * Then created 2 separate queries on the canvas pulled from that initial

Tweets Featured

Teaser - Value Scrubbers

Value scrubbers are done(ish)! The {unnamed} project is shaping up to be a solid UX for data views / pivots in a fluid & creative way - but in a way that retains "low level flexibility" (i.e. retaining full vega-lite, hiccup, re-com, & sql "hackability").#sql #clojure #dataviz pic.twitter.com/