TracEMAIL: Final Project

by nkurani @ 3:00 am 10 May 2011

TracEMAIL attempts to better understand the relationships between time, people, and emails. The user can control a slider to view a snapshot of their email traffic on a selected date. This is achieved by using processing to read gmail messages via thunderbird.

I was inspired by projects that were created by Ben Fry and Fernanda Viegas. Here are some screenshots of visualizations that I impacted my final project:

Here are the links of papers and visualization descriptions that I read through to learn more about how others have parsed similar types of data: (5 different email projects)

For my final project, I decided to revisit data visualizations. I felt as if this is an area where I can grow and really benefit from a second attempt. I decided to explore the realm of email data.

At first, I started by using applescript to extract information from my emails. I created a text file aggregating all the information into a single tab delimited table. I also used it to convert the emails into text files.

Since this process took far too long, I was advised by Golan Levin to instead use Processing to parse the emails in a faster way. He got me started with the scraping of the .elmx files stored by the Mail application on my Macbook. It was tricky because the format of the information in the .elmx file isn’t the same from email to email. This slowed me down in the parsing process; however, I overcame this obstacle by simply revising my code to handle various formats for the same kind of data, i.e., the date.

Once I was able to parse through the files, I experimented with the visualization. I began by simply drawing a square for each email. I made the squares where the email frequency was higher a darker shade. That way emails from frequent senders are highlighted. When you select an email, I had it display onto the screen.

I realized very quickly that this isn’t the best visualization. I had tons of data, and I was now at the point where I wanted to present it in a meaningful way. I experimented different ways to present the information. I also sketched out several of the ideas. To help explore multiple ways to visualize this data. I also read papers for inspiration. The links provided under “BACKGROUND REFERENCES” have various .pdf files of papers, which I read to learn how people have handled email data in the past.

In the end, I was most inspired by Ben Fry’s baseball example, the email mountain, and the frequency of emails. I really liked the clean way Ben Fry was able to compare the salary vs performance of the baseball teams, and the way the Fernanda Viegas was able to visualize a new phase in her life through the creation of email mountains while also using her email data to view how many emails she is sending and receiving from specific individuals. These really inspired me to create my final project.

Through the use of tracEMAIL a user can view how many emails are being sent and recieved from a specific individual. You can track this information over time to see how your relationships have changed. This really combines some of the concepts I learned about while searching for inspiration.

Here is an image of my work at the final show:

Here is a screenshot of the visualization:

Here is a short clip filmed at the final show:
[Coming soon! Vimeo/YouTube currently blocked by hotel!]

I feel that this was definitely a successful exploration of email data. The data was very difficult to parse especially because the .elmx files were so hairy and my computer is so super slow! However, I was able to work past that and create a meaningful, interesting visualization. There is definitely still some room for growth; on my next iteration, I would like to change the thickness of the connecting lines based on the avg length of the emails that are being sent back and forth. I also need to add an indicator to point out that the names that appear on the top of the list are the most frequent senders/receivers of my email. There was definitely some confusion over that. I’d also like to make the scrolling of the slider to view the date a bit smoother. Overall, I am proud of my attempt and final project.

Here is a link to the code:

This work is licensed under a Creative Commons Attribution-Noncommercial-Share Alike 3.0 Unported License.
(c) 2017 Interactive Art & Computational Design / Spring 2011 | powered by WordPress with Barecity