Navya Kondur thesis recap: Using K-Mode Clustering to Identify Personas

Navya Kondur successfully defended her thesis titled Using K-Mode Clustering to Identify Personas for Technology on the Trail on April 19, 2018.  Navya wasn’t originally thinking to do a thesis when she started examining some of the questionnaire results from Gracie Fields’ thesis, but Navya identified some interesting questions within Gracie’s data sets and, most importantly, a new way to examine the data.

Navya presented some preliminary results at the GROUP 2018 workshop, receiving some great feedback from Mike Jones, who has been working on some similar persona creation activities. Navya’s background in statistics served her well in highlighting some possible ways to analyze results from some of Gracie’s data sets; specifically, by using k-mode clustering to identify groups of like-minded hikers.

K-mode clustering is a method to identify clusters within categorical data.  It is a modification of the more popular k-means analysis, adapted for use with categorical data such as types of gear that people bring on hikes, sleeping preferences when on multi-day hikes, and selections between paired “would-you-rather” options (as featured in Gracie’s thesis work).  Navya administered a series of “would-you-rather” questionnaires at various Tech on the Trail events, collecting sufficient data to craft clusters.  Since the clusters are not particularly descriptive or evocative, Navya then crafted five personas that helped to reflect some of the differences among hikers; e.g., younger people embrace technology but lack the money for it.

Navya’s thesis really helped to highlight the possibilities in her line of research, particularly with regard to k-mode clustering and persona identification–though the small number of participants yielded a small number of clusters.  However, her work has resulted in a funding proposal that seeks to identify the nuances between people on trails, and, if funded, we will be on track to publish a large-scale examination of this domain in the future (and maybe we can talk Navya into returning to VT for her Ph.D.)  Until then, check out details about this work in Navya’s GROUP workshop paper and (when released) thesis document.

Navya and her committee (L-R: Scott, Steve, Mike) at her defense

One thought on “Navya Kondur thesis recap: Using K-Mode Clustering to Identify Personas

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s