Y-axis autoscale doesn't choose great subscales

[fgmart-zz]

Check out Project 3190 viewed as bar chart, http://isenseproject.org/projects/3190/data_sets/ . This is a 4th-gr teacher's project about kids' favorite pizza types:

So these will necessarily be integral data, and yet the Y-axis subscale is in units of 0.4 "votes".

Is there any way this can be made more user-friendly?

[VictoriaAlbanese]

Propositions

When setting up fields manually, the user has the option to add a restriction on a string. The restriction column is filled out when the text data type is being used, but is blank otherwise. I propose that we add a checkbox when the data field is a number that can restrict the data to integers. There is a field in the y-axis component of highcharts, allowDecimals (set to true by default) which restricts the axis labels to integers. It doesn't make sense to me to make assumptions about all data, but if users want to restrict their data for certain projects, this makes more sense. This does not change the fact that minor axis lines may seem "non intuitive", but it guarantees that axis labels are. As for the minor axis positions, one could try and implement a function where the minor axis ticks are on whole numbers only (playing with minorTickInterval attribute of the yAxis on highcharts).

A feature which could come from this project which may be generally useful would be something similar t the histogram bin size slider where the user can use a slider to determine what the minorTickInterval should be for their particular dataset.

The latter idea may be something to separate out into a separate bug, where the first suggestion is more of an "MVP" for this particular issue. While most data cannot be restricted to decimals, there is no reason why we can't make the user experience more friendly for "voting" and other applications where decimals just obfuscate things.