People who work in finance make a lot of charts.
Often they chart something, notice an anomaly, and then seek an explanation.
That can involve a lot of searching.
One of the more exciting product challenges we’re working on is how to automatically annotate charts to provide insight.
To illustrate the challenge and opportunity, I included a chart of the volume of an arcane data set: the number of daily news stories about Goldman Sachs since 2005.
There’s an obvious pattern, but its not clear why.
By overlaying icons for the date of earnings releases, it becomes immediately understandable that those announcements drive news coverage.
It also shows the magnitude, ie Goldman gets two to three times the average daily volume of media coverage when it announces results.
The next step is hard, but not a giant leap to imagine using technology to find data sets that explain patterns and automatically apply them to charts.