You know, if you're already on the command line, gnuplot and github already provide ~similar functionality. I can do the same in ~the same number of keystrokes if things are aliased properly. Also, keeping things local is nice.

I use gnuplot to work with large datasets I extract from telco carrier logs (for client usage analysis). It's one of the few tools you can throw a 100,000 point dataset at and it graphs it in seconds. Try that with most scripting language graphing libraries, and you'll be there for a few minutes.

The problem I have with gnuplot -- and maybe I'm the limitation, not the software -- is that I find it difficult to use, and the result of some input is inconsistent. Not to appear self-deprecating, but I really do think I might be the limitation here. Regardless of that, I often waste time trying to figure out why gnuplot is behaving as it is.

I have no idea if chartulo.us is going to choke on my datasets, but I hope I'll be able to use it. I really could use a tool that is a little easier to use than gnuplot.

Tableau does nearly instant plotting of millions of data points with a nice user interface; it may be worth checking out if your job depends on visualizing massive amounts of data.

Hey, that's a good point. We're also interested in collaboration. How do you share your charts once you create them?

Nothing fancy. Either a shared folder in my home-directory for users on the same machine, or email. Most of my work is separate individuals showing off their data rather than working together on one dataset. That could be a cool feature though--project groups or something.

Ahh cool, i see the attraction. I went through a phase of using the google chart API instead of GNUplot (you just wget a URL piping your data). Coupled with a big bash history it works fairly well.

I've learned enough R to be dangerous. Ggplot2 looks gorgeous without trying. It's my current favourite torch for shining on perf and capacity issues (I dream of one day identifying a perf issues root cause with nothing more than the correlation function!)

It looks nice.

If you're looking for an open alternative there is https://github.com/mikedewar/d3py which plays very well numpy/pandas.

Ah, that's a cool project! We will check it out.

I use ggplot2 all day baby. http://had.co.nz/ggplot2/

It has a bit of learning curve but its an extremely flexible way to quickly visualize data in a lot of different ways to get a feel for what's going on in your dataset (you should probably already have some R familiarity, the language has its idiosyncrasies).

Looks cool, I've been wanting something like this for a while.

Also, your animated quickstart is annoying. Leave it as text please.

That's great. Can you give us some idea of what kinds of data you're looking to chart and what sort of ability you'd like out of a tool like this?

Noted on the animation.

Not that it really matters, but

  chartulous < data.csv

http://partmaps.org/era/unix/award.html

Nice, we already won an award! :)

At least from my perspective, a major problem is maintaining control of your information. How does Chartulo.us store the data? Where? Can it be removed immediately post retrieval?

I would much prefer generating the graph locally and then uploading to the service - it's very rare that graphs I generate need to be shared rather than embedded in a paper.

OK. Are current tools (like gnuplot and the others mentioned in this thread) good enough for you or is there some room for improvement?

another option/alternative is to use matplotlib, which is ideally integrated with ipython notebooks http://ipython.org/, well worth a look if anyone's interested.. their integration with numpy/pandas etc is superb.. and on top of all the benefits of plotting with matplotlib, you get ipython and all its features for free... for example http://imgur.com/Le8px

The animation of the command line is a little annoying because it has all this cruft of the package installer messages. I got bored half way through and missed and had to replay. Just put the two command lines there and be done with it. :)

This may be a dumb question but is this a startup or a project?

When (as in the current case) I don't see a way to download source code, I assume it is a startup or a project that will become a startup if it becomes sufficiently popular.

That's right. Right now we're trying to get a handle on the interest in a tool like this.