I’ve just completed and uploaded a project titled Friendship: Not for the Impatient.

The project spawned out of reflection on and appreciation for the relationships I’ve cultivated over two years of college. College poses a serious disruption to the social network of high school graduates, especially those moving far from home. Starting from scratch, one has to relearn the art of friendship cultivation no longer surrounded by the people they grew up with. This presents an incredible opportunity to reinvent oneself free from prior reputation but managing personal growth while finding meaningful companionship is a tall task. This project aims to find insights into the process of college friendship development by answering the following questions:

1. From two years of messages in a group chat of 31 close friends, can we extract an individual measure of the friendship strengths between pairs of members?

2. What can we learn visually from graphically analyzing those relationships over the time span?

3. How many months of interaction are needed to accurately predict the two-year outcome of a friendship?

Read the article for further methodology and explained results and the guide for details on how to access and wrangle your own iMessage data.

I’ve spent a lot of time reflecting both on paper and mentally about my own development over the past two years and the results from this project validated many of my hypotheses. The dataset, which will remain confidential for the sake of its member’s privacy, offers real insight into the college transition process both for my friends and for teenagers as a whole.

It’s a project that consumed most of my limited free time over the past six months and that offers many future opportunities for further analysis. Specifically, in the coming months I hope to employ techniques I learned in the NLP course I took this past semester beginning with unsupervised conversation disentanglement as well as conversation summary and group sentiment analysis.