- Identify gaps by null values in series. Null value = don't draw column.
- After not drawing individual columns in a cluster, equalize distance between clusters. Adjust column widths accordingly.
- Control this behavior with flag: Hide null columns.
This is a frequent issue for us. If we want to show a 'Top 10 Cities per Continent' we run into the need for workarounds for two problems: cities are unique to each continent, and some continents have <10 cities (in our dataset).
In slightly more complicated scenarios, we may want a 'Top 10 Products per Continent' graph, in which some products are in the top 10 for multiple continents, while others are unique to 1 continent. Currently this would result in [count distinct products across all continents] x [# of continents in dataset] columns, when a maximum of 10 x [# of continents in dataset] is desired.
Milan Gamble
supported this idea
·
Grzegorz's solution is close. What's left:
- Identify gaps by null values in series. Null value = don't draw column.
- After not drawing individual columns in a cluster, equalize distance between clusters. Adjust column widths accordingly.
- Control this behavior with flag: Hide null columns.
This is a frequent issue for us. If we want to show a 'Top 10 Cities per Continent' we run into the need for workarounds for two problems: cities are unique to each continent, and some continents have <10 cities (in our dataset).
In slightly more complicated scenarios, we may want a 'Top 10 Products per Continent' graph, in which some products are in the top 10 for multiple continents, while others are unique to 1 continent. Currently this would result in [count distinct products across all continents] x [# of continents in dataset] columns, when a maximum of 10 x [# of continents in dataset] is desired.
Example: http://imgur.com/a/ErH17