Large variation in y-axis values

[bigroi]

I am trying to chart data with a large variation in values. Example, data from January-December with values ranging from 150,000 to 6.5 million.

When I chart this, the y-axis scale is so large that you can't tell anything about the small month values.

Any help in correcting this problem would be greatly appreciated.

[MrShorty]

Jon Peltier (www.peltiertech.com) describes how to do a broken y axis somewhere on his site. Will that work for you?

[Carim]

Hi,

A great solution is provided by Andy ... with a broken Y axis ...
http://www.andypope.info/charts/brokencolumn.htm

[bigroi]

That was quick...thanks friends! I'll check out those sites.

[MDubbelboer]

I actually really like that broken y-axis stuff and have never seen that before and am interested in trying it out, but the distribution of your data may not be suited for that.

if your data only has one value at either of the extremes i'd just throw it out as an anomaly.
if your data has multiple values at both extremes then the broken y-axis method is suitable
if your data is uniformly distributed across the range then a transformation may be required, something like a log transformation will standardize your data around a much smaller distribution, however you'd have to be aware of what this transformation is doing to your data and realize that it cannot be read in the same manner as non-transformed data