3 ms·

This graph is about daily domain changes. Maybe if you looked here, it would help (the four values in the four tab headings): http://www.dailychanges.com/domain

by zeratul 15y ago

This graph is about daily domain changes. Maybe if you looked here, it would help (the four values in the four tab headings): http://www.dailychanges.com/domaincontrol.com/ http://www.dailychanges.com/domaincontrol.com/ I'm experimenting and leaning new ways to visualize the data. I think there are better ways of doing it than a bar graphs or line plots. Well, it seems it did not work that well. GitHub uses similar visualization technique, though. Maybe next time I get it right. Here is my inspiration: http://raphaeljs.com/github/impact.html http://raphaeljs.com/github/impact.html EDIT: Initially I wanted to do Sankey diagram but that came out way too complicated and there wasn't a good way to show new and deleted domains.

ivankirigin 15y ago
Stacked graphs are hard to compare to eachother. Eye-balling the height of middle veins of data is especially hard. Comparing a line graph is easier because the height you're estimated is the difference between values, and you already know which is absolutely larger. Two specific problems with your visualization are the margin between the veins and the fact that you scroll horizontally. This data is very much like tracking the active users on a service, with new, churned, retained, and resurrected users. I'm very familiar with that topic and think your visualization is bad for this data.
stephenatwork 15y ago
One problem I have with the graph is that the height of the sections of data for each day does not have a clear meaning. For instance, look at Dec 23 and Dec 24: the top orange section drops from 31k to 29k, but the visualization shows a thicker orange height on the 24th than the 23rd. In fact, the entire graph is significantly higher on the 24th, but it is not clear what this means either.
- zeratul 15y ago
  I've updated the graph according to your suggestion. Before every variable was normalized according to its own min/max, now there is a global min/max. Thank you all for the suggestions.

by zeratul 15y ago

ivankirigin 15y ago

Stacked graphs are hard to compare to eachother. Eye-balling the height of middle veins of data is especially hard. Comparing a line graph is easier because the height you're estimated is the difference between values, and you already know which is absolutely larger. Two specific problems with your visualization are the margin between the veins and the fact that you scroll horizontally. This data is very much like tracking the active users on a service, with new, churned, retained, and resurrected users. I'm very familiar with that topic and think your visualization is bad for this data.

stephenatwork 15y ago

One problem I have with the graph is that the height of the sections of data for each day does not have a clear meaning. For instance, look at Dec 23 and Dec 24: the top orange section drops from 31k to 29k, but the visualization shows a thicker orange height on the 24th than the 23rd. In fact, the entire graph is significantly higher on the 24th, but it is not clear what this means either.

zeratul 15y ago
I've updated the graph according to your suggestion. Before every variable was normalized according to its own min/max, now there is a global min/max. Thank you all for the suggestions.