Leader Lines: Even the Simple is Befuddling Me!
Hello! Please see the attached. These are point features on the roads; I have a no color symbol engaged. The balloon one is me fiddling w/annotation and reflects the look I'm going for. However, there will be a couple hundred of these little intersection maps within this project, so an automated solution is highly desired! I think simple callouts would better suit the application, but look at how lousy my best effort looks. Despite the roads having a super high weight (1000) no combination of leader tolerance/offset prevents the numbers from conflicting with the linework. Can I do better? Thanks! (ArcMap10SP1, Maplex)
Mapping Center Answer:
Okay--first I think there's a bug here because what you had, with one change should have worked. That change was to turn off the Never Remove Labels setting.
I made five other changes that produced a better result (the first 3 had little consequence):
1. Reduced the feature weights from 1000 to 800 (this helped with labeling the linear features).
2. Changed the maximum offset to 500% (no point in doing 3500% -- you could get a really ugly label).
3. Optimized your label expression to have two fewer lines:
Function FindLabel ([STANUM])
FindLabel = [STANUM]
IF InStr([STANUM], "-") <> 0 THEN
FindLabel = INT(Right([STANUM],4))
4. The big one was to adjust how you were using user defined zones. I just used the 3:00 and 9:00 positions--giving them a rank of 1 and 2 respectively. This forced many of the less good looking labels off the map (I noted that most of the collisions occurred in close to vertical orientations).
5. I changed your point symbol to be a 25pt circle (using the circle symbol from the default style). Then I gave that a label weight of 800 as well.
I'm not totally happy with the result, but it's much better than before. The only issue is that Maplex often insisted on crossing a road with a leader--doing #5 alleviated some of that.
I gave this case to the development team to have a look at.
If you would like to post a comment, please login.