8 ms·
HiPlot: High-dimensional interactive plots made easy
- juliwu 7y agoDoes anyone know if this can be integrated with tensorboard.
- sillysaurusx 7y agoActually yeah, I wrote a script which 1. exports data from tensorboard.dev, then 2. prints it out in csv form that can be dropped into hiplot. It was ok. Kind of lost interest when I saw the actual result. You can't smooth the data, so often times it's hard to tell what's going on. Here's the awful script. It's specific to my own naming conventions, so it probably won't work for you. But the same idea would work: just parse the tensorboard data and spit it out as csv. https://gist.github.com/shawwn/b74d6e58da6496e2ade02bab61acc920 https://gist.github.com/shawwn/b74d6e58da6496e2ade02bab61acc... I also ran into a bug where multiple different datapoint types were getting merged into one. (It wasn't due to the csv code; the actual CSVs are fine.) I.e. the rows in the csv had a certain column that had a certain value, but in hiplot that column was nowhere to be found. Here's what it looks like: https://imgur.com/1n1eDXu https://imgur.com/1n1eDXu I deployed an example here: https://hiplot-subsim-demo.now.sh https://hiplot-subsim-demo.now.sh Notice that the "run" column shows up in the data, but not in hiplot.
- rckoepke 7y agoReminds me of some charts used by chemical engineers in the age of slide rules. Can anyone help me remember an example?
- chemeng 7y agoDon't know if it's what you're referring to, but psychrometric charts are always fun, and any of the thermodynamics charts (Pressure-Enthalpy being an example). The best part? You say the "age of slide rules", but these are still used today.
- rckoepke 7y agoNot psychrometric...the ones I'm thinking of look exactly the same as this, but allow you to solve a system of equations by drawing a straight line between points on two columns, thus making the intersection of the line with the other columns your "answer" for those variables. The scales are sometimes very strange, in order to achieve this effect.
- Cybiote 7y agoThese are known as https://en.wikipedia.org/wiki/Parallel_coordinates https://en.wikipedia.org/wiki/Parallel_coordinates and are a pretty old technique. Although the wiki lists process control as one of the major users, their history has them as dormant until revived in their modern form in the 1960s, which doesn't quite match your timeline. Perhaps the discrepancy is resolved by whether parallel axes were used as a equation solving aid versus a visualization aid? They're also not outdated, you can find them in many libraries, like highcharts here: https://www.highcharts.com/docs/chart-and-series-types/parallel-coordinates-chart https://www.highcharts.com/docs/chart-and-series-types/paral... Excellent technical coverage is: Parallel Coordinates for Visualizing Multi-Dimensional Geometry, from 1987 at https://link.springer.com/chapter/10.1007/978-4-431-68057-4_3 https://link.springer.com/chapter/10.1007/978-4-431-68057-4_... Wikipedia links to this 1880 chart: https://www.davidrumsey.com/luna/servlet/detail/RUMSEY~8~1~32803~1152181:General-summary,-showing-the-rank-o https://www.davidrumsey.com/luna/servlet/detail/RUMSEY~8~1~3...
- rckoepke 7y agoThe ones I'm thinking of were "nomograms" (thanks to user 'chemeng'). Visually they look like perhaps a very specific subset of "Parallel coordinate" graphs. Definitely more for calculation rather than visualization. Evidently they were popular from around 1890 to at least 1950.
- kortex 7y agoThey are still quite useful for predicting boiling points under reduced pressure. https://www.sigmaaldrich.com/chemistry/solvents/learning-center/nomograph.html https://www.sigmaaldrich.com/chemistry/solvents/learning-cen...
- keanzu 7y agoWe first chose to display only data points obtained after 20 or more epochs of training. Then, by slicing through the “loss” axis, we observed that larger learning rates led to better performance (perplexity). You can reproduce this example here: https://facebookresearch.github.io/hiplot/_static/demo/ml1.csv.html?hip.color_by=%22valid+ppl%22 https://facebookresearch.github.io/hiplot/_static/demo/ml1.c...
- mbostock 7y agoBased on a comment buried in the source, this library seems to be heavily based on work by Kai Chang: http://bl.ocks.org/syntagmatic/3150059 http://bl.ocks.org/syntagmatic/3150059 It’s a shame Kai isn’t created in the README, LICENSE, or announcement.
- gavinray 7y agoAnd Kai credits previous work from Jason Davies[0] and Mike Bostock[1]. Mike Bostock is the creator of D3.js and ObservableHQ. [0] https://bl.ocks.org/jasondavies/1341281 https://bl.ocks.org/jasondavies/1341281 [1] https://observablehq.com/@d3/parallel-coordinates https://observablehq.com/@d3/parallel-coordinates Edit: Just saw your name. Holy shit, it's Mike Bostock. On a side tangent, I find it crazy that the New York Times for a long period had yourself, Jeremy Ashkenas, and Rich Harris all on staff. What an all-star team.
- kovek 7y agoMike Bostock (mbostock in this thread), D3.js is awesome!
- mstade 7y agoYou may already be aware of this, in which case I apologize for stating the obvious, but Mr. Bostock actually authored the comment you replied to. :o)
- gavinray 7y agoI about jumped out of my chair when I saw the username after posting
- mbostock 7y agoJeremy’s here at Observable, too!
- gavinray 7y agoHuge fan of you both. Best wishes, and I get to go home today and tell the missus I engaged in unknowing dialogue with Mike Bostock, about Mike Bostock.
- zetaben 7y agoFun fact, in the network and security world, there used to be a tool called picviz that was doing exactly the same kind of things. https://doc.ubuntu-fr.org/picviz https://doc.ubuntu-fr.org/picviz (Sorry I could only find it in French). Seems defunct nowadays.