Comments on: Axis Labeling http://www.justintalbot.com Mon, 31 Oct 2011 14:58:34 +0000 hourly 1 http://wordpress.org/?v=3.2.1 By: Justin Talbot http://www.justintalbot.com/research/axis-labeling/comment-page-1/#comment-409 Justin Talbot Mon, 31 Oct 2011 14:58:34 +0000 http://www.justintalbot.com/?page_id=154#comment-409 Thanks, I'll check the gnuplot code. I haven't used this for log scales, but using the exponents (e.g. -12 to -6) should provide plausible, if not optimal, results. You're right that for log scales, non-integer numbers are much less nice than for standard scales. You would have to adjust the niceness score to account for this. Thanks, I’ll check the gnuplot code. I haven’t used this for log scales, but using the exponents (e.g. -12 to -6) should provide plausible, if not optimal, results. You’re right that for log scales, non-integer numbers are much less nice than for standard scales. You would have to adjust the niceness score to account for this.

]]> By: Justin Talbot http://www.justintalbot.com/research/axis-labeling/comment-page-1/#comment-407 Justin Talbot Mon, 31 Oct 2011 14:53:08 +0000 http://www.justintalbot.com/?page_id=154#comment-407 The simplicity component of the optimization score includes a term for whether or not the labeling includes 0. You can increase the weight on this term if you like. The simplicity component of the optimization score includes a term for whether or not the labeling includes 0. You can increase the weight on this term if you like.

]]> By: Claudia http://www.justintalbot.com/research/axis-labeling/comment-page-1/#comment-406 Claudia Thu, 27 Oct 2011 12:11:37 +0000 http://www.justintalbot.com/?page_id=154#comment-406 Hi: In hopes to get some insight, I would like to know if there is a way to force the zero value to be an axis label. I have found that in some types of charts it looks weird if the zero value along the y axis is not displayed. This is the case for the bar chart, amongst others, specially if you have both positive and negative values. I know it is possible to tweak the settings, but this will depend on the data ranges being assigned. If anyone would like to share information on how this could be achieved predictably, independent of the other settings, I would be immensely grateful. Hi: In hopes to get some insight, I would like to know if there is a way to force the zero value to be an axis label. I have found that in some types of charts it looks weird if the zero value along the y axis is not displayed. This is the case for the bar chart, amongst others, specially if you have both positive and negative values.

I know it is possible to tweak the settings, but this will depend on the data ranges being assigned. If anyone would like to share information on how this could be achieved predictably, independent of the other settings, I would be immensely grateful.

]]> By: christie harper http://www.justintalbot.com/research/axis-labeling/comment-page-1/#comment-404 christie harper Wed, 19 Oct 2011 20:50:32 +0000 http://www.justintalbot.com/?page_id=154#comment-404 very interesting work. i have implemented the majority of your stuff in matlab and if interested will send you a copy (each method extended, wilkinson... are separate functions) if you are interested. minor bug in the gnuplot portion code (at least the R), towards the bottom where graphmin & graphmax are set you have the values min & max vs dmin & dmax. I have see numerous references to R before but this is the first time i have played with it. have you attempted to apply any of this to log scales? (i simply tried each one and using the actual values 1.0e-12 to 1.0e-6 did not work and trying -12 to -6 sometimes gives me non integer exponents). thanks very interesting work. i have implemented the majority of your stuff in matlab and if interested will send you a copy (each method extended, wilkinson… are separate functions) if you are interested.
minor bug in the gnuplot portion code (at least the R), towards the bottom where graphmin & graphmax are set you have the values min & max vs dmin & dmax. I have see numerous references to R before but this is the first time i have played with it.
have you attempted to apply any of this to log scales? (i simply tried each one and using the actual values 1.0e-12 to 1.0e-6 did not work and trying -12 to -6 sometimes gives me non integer exponents). thanks

]]> By: Justin Talbot http://www.justintalbot.com/research/axis-labeling/comment-page-1/#comment-368 Justin Talbot Wed, 31 Aug 2011 18:07:13 +0000 http://www.justintalbot.com/?page_id=154#comment-368 Here's <a href="http://www.justintalbot.com/wp-content/uploads/2011/08/AhmetJava.zip" rel="nofollow">Ahmet's Java code</a> which he's given me permission to post. I haven't looked at it too carefully, yet. If you run into any problems, let me know and I'll pass the message to Ahmet. Here’s Ahmet’s Java code which he’s given me permission to post. I haven’t looked at it too carefully, yet. If you run into any problems, let me know and I’ll pass the message to Ahmet.

]]> By: Hamid Younesy http://www.justintalbot.com/research/axis-labeling/comment-page-1/#comment-367 Hamid Younesy Mon, 29 Aug 2011 20:09:19 +0000 http://www.justintalbot.com/?page_id=154#comment-367 so... the answer is yes? :) so… the answer is yes? :)

]]> By: Hamid Younesy http://www.justintalbot.com/research/axis-labeling/comment-page-1/#comment-365 Hamid Younesy Wed, 17 Aug 2011 12:31:01 +0000 http://www.justintalbot.com/?page_id=154#comment-365 Hi, I am interested in using your labeling approach for our visualization framework which is in Java. I was wondering if there is a Java implementation might be available sometime soon, or I should try to port the C# version myself. Also wondering if Ahmet Karahan has finished his Java implementation and if he is willing to share it. Thanks! Hi,
I am interested in using your labeling approach for our visualization framework which is in Java. I was wondering if there is a Java implementation might be available sometime soon, or I should try to port the C# version myself. Also wondering if Ahmet Karahan has finished his Java implementation and if he is willing to share it.
Thanks!

]]> By: Justin Talbot http://www.justintalbot.com/research/axis-labeling/comment-page-1/#comment-340 Justin Talbot Thu, 16 Jun 2011 16:44:00 +0000 http://www.justintalbot.com/?page_id=154#comment-340 Inf is just infinity. In C, you could either use the true floating point infinity, or just set it to a really large value like you did. The break conditions are used to shortcut the search. It's pretty straightforward to prove that if the break condition succeeds that any further labelings generated by that loop would be worse than the best one found so far. So we quit out of that loop and go on to the next iteration of the containing loop. Id be glad to see your code. Inf is just infinity. In C, you could either use the true floating point infinity, or just set it to a really large value like you did.

The break conditions are used to shortcut the search. It’s pretty straightforward to prove that if the break condition succeeds that any further labelings generated by that loop would be worse than the best one found so far. So we quit out of that loop and go on to the next iteration of the containing loop.

Id be glad to see your code.

]]> By: Justin Talbot http://www.justintalbot.com/research/axis-labeling/comment-page-1/#comment-339 Justin Talbot Thu, 16 Jun 2011 16:41:39 +0000 http://www.justintalbot.com/?page_id=154#comment-339 Real soon, I promise. Real soon, I promise.

]]> By: Justin Talbot http://www.justintalbot.com/research/axis-labeling/comment-page-1/#comment-338 Justin Talbot Thu, 16 Jun 2011 16:41:18 +0000 http://www.justintalbot.com/?page_id=154#comment-338 Tim, the dataset was randomly generated using code that you can find the in R library that's been released. Tim, the dataset was randomly generated using code that you can find the in R library that’s been released.

]]>