Another cool new feature launched tonight. Every field in every datablog now has six links auto-generated to graphs for various time periods. This includes custom fields that you add to an existing log type or custom fields that you add to your own log type.
We've had graphs of data fields for a long time... it's so 2003. But for a long time we've had them somewhat hidden in the UI. By putting a set of links on each field, right in each blog entry, we make it easier to get to the graphs.
For example, you're reading my blog and you see that yesterday I ran 14.1 miles around San Francisco. "Is that typical mileage for this assclown," you ask yourself? Yesterday you'd have had to dig to find the answer. But not today. No siree Jim Bob! By simply clicking on one of the links you're shown a graph with my historical performance in that field... in this case, running distance.
The time periods we're linking to are:
This Week | This Month | This Year Last 7 Days | Last 30 Days | Last 365 Days
We chose these as the most common that people would want to see. Check out the screenshot... you'll see that this adds a lot of links to the screen. We may want to do some more dhtml hiding at some point... maybe a small icon which, when moused over, would scroll out the list of available graphs for that field.
(Ooooooh, or a nice little ajax-powered dhtml window that creates a small version of the chart when you mouseover the link. Ooooh. Ahhhh. Not too hard to do, but the server load might be a bit high for me right now. Hmmmm... gotta noodle that one. A caching mechanism for graphs may be in order.)
if the field is numerical or time-based the link takes you to that field vs. time. If the field is alphanumeric the link takes you to a pie chart representing the number of times that an entry has been made with that value.
An alphanumeric example may be something like my shoes in a running log. By clicking the This Year link you're shown the percentage breakdown of the number of times I've run in each pair of shoes over the course of the year. Snazzy!
Also, keep in mind that once you're on the graph page you can customize the x-axis, y-axis, chart type, time period, chart size, y-axis aggregation (sum or average), etc. You can also, and this massive power is always overlooked, apply a graph to a saved search.
Applying a graph to a saved search is massive power. An example: I create a saved search to only see entries from my running log where the distance field is over 13.1 miles, a half marathon. Bammy... a list of entries pops out and you can one-click that search in the future. Cool in and of itself. But now, by applying the graph to the saved search (essentially a dynamic sub-set of blog entries) you can gain a bunch of new insight. Which shoes do I generally wear on longer runs? What's my average pace when I run over 13.1 miles? Etc. I can slice my data and then graph it. Powerful... and almost always overlooked... which is a UI issue, not a user issue... we're working on it.
Opening the archive. Personal data mining. That's what we do. Let us know what you think.
Tuesday, December 5, 2006