Have you ever wondered what your e-mail says about you? Conversation is ephemeral; e-mail leaves a record. The e-mails you send and receive and the meetings you have tell stories. Now, there is fun software that extracts those stories for you. And I can see how the underlying code and concept could be helpful in e-discovery.
Cataphora recently released Digital Mirror, which “is designed to help you understand your behavior and those of others in your digital ecosystem.” This software is released in connection with Cataphora CEO Elizabeth Charnock’s new book about the Digital You.
The company explains it has applied its “software and expertise to uncover evidence in major litigation and investigations” along with “patented data analysis technologies” to create this software, which “allows you to see a reflection of your digital universe and your role within it.” For now, the free download works only on your own Outlook files; future versions will do more. I’ve installed and run the software here. I share here my experience and comments.
Installation
The download is free (here). Installation ran smoothly. The first step is indexing your Outlook messages. Doing so is easy; the software automatically found all of my PST files. As warned, however, the indexing takes time. Mine took about 2 hours and I was able to use my PC during indexing.
Digital Mirror indexed 70,000+ items for me. I have PST e-mail archives that go back to 2003, so that’s a lot of years of e-mail. Nonetheless, the count seemed high since X1 indexed only 58,000. The Digital Mirror FAQs explain that it indexes events (appointments) and contacts. Adding those in, the count I would expect is fairly close to what was indexed. Once the indexing was completed, I could not find a way to see which PSTs were indexed nor how many items.
The Stories: What Digital Mirror Tells You
The Digital Mirror software opens with a dashboard displaying several tools to learn about the digital you. I discuss a few in detail:
Quality Time. “The Quality Time visualization gives you some insight into who you devote your valuable time to. Digital Mirror’s analysis draws on Outlook information about meetings you attend, along with how many emails you write – especially long ones that typically require more time and care to compose.” The website provides a top level explanation of how these results are derived. For example, the software uses semantic analysis to determine if e-mail messages are about meetings and gives these more weight than just e-mail.
The results are presented as a pie chart showing the “Top 5” people. The pie chart changes every month and a time line at the top lets you look at one month or “play” the time line as an animation, which lets you can see how your Top 5 evolve month by month.
Playing the time line was a trip down memory lane for me. The pie slices showed names of people with whom I met or corresponded. I saw few surprises but was reminded both of big work projects and some periods of intense e-mail conversation with friends or family.
I think this is a good and simple display, easily understood. I look forward to a future version that shows more people and adds feature. For example, it would be nice to be able to control the speed of the animation or to show two pie charts at once from any two arbitrary months.
Social You-niverse measures “closeness” to contacts. “Closeness is measured using a wide variety of factors, including how you greet each other when writing emails (Dear Professor Einstein versus Hey Al! for example), as well as analysis of emotive tones used in emails between you. ” The display is solar system with you in the center and contacts orbiting you at very distances. It too uses a time line, which you can play as an animation.
I found a a fair bit of overlap here with “Quality Time”. It’s hard for me to tell the difference between “quality” and “closeness”.
Who? What? When?. This shows a grid of “hot topics” and who participated in the conversation. Hot topics are determined by linguistic analysis that includes how often you reply, e-mail length, “Whether you tend to store messages about the topic” (though if not stored, the software would not know about the topic so, hmmm?), and “Whether the topic is associated with a ‘good’ group of correspondents.” In this display, you can click on column heads, row stubs, and the cells themselves for additional information. Having more “drill down” features like this will be crucial for commercial grade software.
This is potentially the most interesting analysis but I found it very limited. Digital Mirror selected five topics across several years. (Interestingly, the analysis here only begins in 2007 – the others begin in 2003.) Five topics over 3+ years just does not tell me very much. And one of those topics “Google Alerts” is a red herring. I do forward a lot of Google Alerts but that is for their content and links. This shows the challenge of computer-based linguistic analysis.
Other Features. Digital Mirror also includes displays for
– Blow Off Scoreboard
– Pecking Order
– Buck Passing
– Loud Talking
– Temperature Gauge
– Stressful Topics
I found these less useful / interesting in the current version though they all seemed accurate as far as I could tell. I can see, however, where this type of analysis could be interesting. Note that in some displays, a mouse click in a matrix brings up a menu to advance or go back a topic.
Comments on the Visual Display
I find the analysis and some of the measures very interesting but the visual display could use some work. Backgrounds are too busy / distracting for my taste and both the dashboard and data display icons are, in my opinion, corny. The real work here, however, is the analysis and I suspect cleaning up and making the displays more business-like would not be a big task.
Implications
This is the first product I know of (or at least remember!) to bring semantic analysis to the desktop. Of course desktop search makes full-text indexing readily available. But Cataphora goes much further in using linguistics to derive meaning from all the words – and the metadata – stored on our hard drives.
For purely personal use, I suspect Digital Miror will remain more in the category of fun than genuinely useful. For e-discovery and investigators, however, I can see how the underlying technology could be very helpful. If I wanted to get a quick and easy sense of what someone worked on and who they corresponded with, I think the approach (if not all the actual displays) here would be very useful. Of course, this is not surprising since Cataphora had its start with e-mail analysis (see, e.g., my 2004 blog post, Using Visualization to Manage Legal Organizations, which mentions Cataphora.)
We see more and more articles about our digital footprint and how hard it is to eradicate. Today, it is hard to get a clear picture of what all the data say. Digital Mirror represents a very good first step in being able to understand the data. Being able to expand this approach to cover social media and websites, as well as the desktop, would be interesting indeed.