netj / 3x Goto Github PK
View Code? Open in Web Editor NEW3X — a Workbench for eXecutable eXploratory eXperiments
Home Page: http://netj.github.io/3x/
3X — a Workbench for eXecutable eXploratory eXperiments
Home Page: http://netj.github.io/3x/
Instead of needing to click on an SVG element, show popovers on hover. If a user clicks on an SVG element, do not hide the popover.
This also entails adding invisible, event-handling rectangles for each visible bar in the bar graph to accept the mouseover event. The main purpose of this is to alleviate the problem of zero-height bars that would normally not allow the user to see the popover.
For bar charts, which show ordinal/nominal values on the x-axis, don't show all of them when there are too many to show because the x-axis labels run into each other. Be able to scale number of ticks according to the width of the screen.
Under the current framework, is it possible to send all planned jobs to servers with a dedicated job scheduler? The workflow would likely be as follows:
Right now, if you have a bar chart selected and plot a quantitative variables vs. another quantitative variable, the x quantitative variable is plotted as a nominal variable. Instead, switch the chart type to a scatterplot and force the user to explicit select bar chart if they want this quant-to-nominal transformation to happen. This is simply the best default to show 2 quantitative variables in a chart.
When adding additional series, bars seem to be offset in a strange manner.
Hello,
I've been running an experiment that takes approx. 4 hours to execute and produces approx. 90000 output lines.
When the experiment is finished, it is shown as Aborted and not as Complete. Looking in the run overview, I can see that 5 of the experiment outputs produce the error "cat: write error: Broken pipe".
I wonder how I can overcome this error and if it is caused by 3x or by my OS (Ubuntu Server 12.04).
In the drop-down that shows additional variables that can be plotted beyond the initial variables for the x-axis and y-axis, filter out ratio variables that don't have the same unit as the y-axis variable. Make sure unitless variables are able to be plotted on the same chart (whose units might be denoted by "" or undefined).
It would be incredible useful to have some additional run information recorded to ensure results are reproducible. Sumatra, does a good job at deciding which details to log.
In particular, storing the SHA-1 for input and config files would help identify modifications from the first execution to first reproduction attempt. Another feature needed for reproducibility, is storing the git head (if the executable code lives in a git repo).
Allow a user to switch between chart types when a set of nominal/ratio variables can be displayed with a different type of chart.
For example, if an input variable is mapped to the x-axis, it could either be interpreted as a nominal or a ratio variable. The former would yield a bar chart; the latter would yield a scatterplot.
Enable the origin toggle for bar charts and set this as the default so that all bar charts are grounded at 0 by default.
Remove in-chart series labels in favor of colored dots next to each axis picker that expose a hidden legend when hovered over or clicked.
For pivoting (creating series), only allow for nominal output variables or input variables. Input variables selected for the x-axis can be treated as either nominal or ratio. If selected for the y-axis, they are treated as ratio. If a non-axis variable (which means they are at least the 3rd selected variable), they are treated as nominal and can cause additional series to be created.
I'm looking for a good tool to benchmark ELKI ( http://elki.dbs.ifi.lmu.de/ ) clustering performance across parameters.
The problem is, that the parameters aren't as nicely uniform as in your examples, and they have strong interdependencies.
The most interesting parameter obviously is the clustering algorithm. Say I'm looking only at k-means and DBSCAN for this example (but there are tons more in ELKI, which is why I could need benchmarking tool support).
The big challenge here are the dependencies of the parameters. The most simple one is that the "k" parameter only exists for k-means, whereas for DBSCAN one needs to choose distance function, minPts and epsilon. But then, there are also k-means initialization heuristics that have parameters such as the random seed...
Will 3x be able to handle such complex cases?
Right now, we choose either an ordinal or linear scale based on variable types. Instead, decide on a scale based on the chart type, since we'll often times want to use quantitative variables as nominal/ordinal ones.
If you choose an x and y, the 3rd axis picker will filter only variables with the same unit as the y variable chosen. However, once you select a 3rd variable, the 3rd variable dropdown menu no longer filters, and instead incorrectly shows many variables to choose from.
Right now, if the x-axis shows rows from the results table with identical names, there will be a separate label for each name. Instead, we need to use a grouped bar chart, where identically-named rows are only once on the x-axis, with no horizontal space between bars.
3x define input 'algo' 'bubble Sort'
gives an error
input/algo=/bubble Sort: Invalid character in pattern value bubble Sort for algo.
It's quite a restriction to have only alpha numeric characters in the input values, unless I am missing some way to have input values with space/comma.
Enable logged bars for bar charts, taking into account the same logic (not allowing log if there are values of 0 in the data) that was employed for line/scatterplots.
Already fixed the problem for scatterplots when quantitative variables are strings, but we need to fix it for all chart types.
3x doesn't take care of the case when it fails the assembly process, and does not report as "Fail" state. 3x remains to report the state as "Running".
Hello.
I would like to know if there is a way to delete past runs, or re-initialize an experiment. I tried removing the runs from the run/ directory but with no luck.
Thanks,
Michalis
Right now, adding additional variables via the dropdown limits only to quantitative variables with the same unit as the y variable. Instead, allow other quantitative variables that will be plotted as nominal variables, as long as that variable has 8 or fewer unique values in the data.
Remove the Y2 axis from being a possibility in the code logic.
The files in the upper directory(inside the 3x storage but not in the input or program directories) are not copied using the "rsync". There is a collision between the --relative and --copy-unsafe-links options.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.