Changing the x-axis crossing point on a log y-axis

You should be able to change the axis minimum in the axis tab of the chart inspector. The min is where it crosses the X axis.

I see. Interesting effect on a bar chart. Area charts too. I see no way around it other than forgo the bars and go with lines.

I have a workaround if you are willing to do some work.

1) Plot it as a line chart. Set the Y axis to log and the minimum to what you want it to be.

2) Set the data symbol to none.

3) Add an error bar, negative only, percentage, 100%.

4) Go to the Graphic Inspector.

5) Change the stroke of the error bar line to be straight, not an I-beam

6) Increase the width of the error bar. Change the color.

7) Select the connecting line. Set the stroke to none.

Will that work for you? I think this procedure needs a little tweaking still. The first series spills over into the Y axis and the last spills off the end of the X axis. Maybe if you plot one blank data point as the first and another as the last it might look better.

Addendum: This only works for a single series, can't do multiple series. Also, if the X-axis labels are not lined up under the bars, add some spaces to the first (blank) label in the Table Inspector.

Message was edited by: Badunit

I have another idea for you.

1) Multiply all of your numbers by 10 or whatever factor of 10 (i.e., 100 , 1000, etc.) is necessary so that none of them are smaller than 1.

2) Take the LOG of those new numbers.

3) Plot the result of (2) using a linear Y axis. Get your chart looking like you want it with the min, max, and number of Y axis gridlines you want. Each number on the Y axis needs to be an integer, no decimals.

4) Choose a custom format for the Y axis numbers. Delete the format that is there and add a new format for each of the numbers on the axis. For example, lets say you multiplied by 100 in the first step above. The gridline with the number "0" should really be "0.01" so make a format for "if value is 0" and type in 0.01 as the format. Repeat for the other numbers (1 becomes 0.1, 2 becomes 1, 3 becomes 10 ,etc).

Again, kind of a pain. I think I like this solution better, though. The chart looks better in the end and you can do multiple data series on the same chart.

I'm still refining. I put in an unnecessary step in the last iteration. You can skip step 2 (using LOG function) and plot the results from step 1 on a log Y axis scale then do step 4.