Andrew Wheeler is a member of the class of 2020 at Hamilton College, studying mathematics and French. His main interest is mathematical analysis and modeling, and he is a founding member of the Hamilton College Data Science Club. Outside of academics, Andrew is an avid trombone player, performing in a wide range of ensembles including jazz combos, wind ensembles and full orchestra. Additionally, Andrew is a member of the men’s varsity cross-country, indoor and outdoor track and field teams.
Project: Visualizing Your Inbox
The tools in the Wolfram Language’s email connection paclet can be very easily used to present and visualize meaningful information about your inbox, including who emails you, who you have conversations with, most common meaningful words and more. The package itself is experimental and very interesting in how it treats each item in any given mailbox, so discovering how each element could be searched and how its details could be presented ended up not being a trivial task.
Main Results in Detail
Many analytical methods transitioned very well to the MailItem entity and, in doing so, clearly explained some very interesting conclusions about how I use my inbox. You typically don’t realize quite how many spam or advertisement emails you get on a daily basis, not to mention various alerts from social media platforms. In the end, a lot of analytical ideas I would have liked to explore proved tough to access as some important pieces of information were missing or algorithmically very difficult to sort and store and even more so to generalize to any given user’s mail account.
Further exploration within the MailItem and MailServerConnection functions would definitely include generalizing functionality to all folders of a user’s mail account and generalizing to work on any given mail servers, as each different provider (Gmail, Yahoo, etc.) uses slightly different tags for a message. More specifically, it would be very interesting to find and group all emails from a specific group of recipients. Analysis on larger conversation groups could show very interesting results, including who gets left out of conversations most often, how quickly certain groups reply and which groups tend to add more recipients as the email chain continues via CCs and forwarding.