/Experience_Project

Kanjoya research project, analyzing the social media network, "Experience Project"

Primary LanguageJupyter Notebook

CDIPS_2015

Summer group research project with mentor Adam Kalman from Kanjoya.

This project is the result of my participation in UC Berkeley's Career Development Initiative for the Physical Sciences (CDIPS) grad student-organized summer data science workshop.

The annual 3 week workshop pairs tech mentors with a small groups of students to work on real industry data science projects. It's a great opportunity for PhDs transitioning into tech to learn the tools of the trade, visit work spaces, and connect with experts.

Our Group

  • Leader: Adam Kalman, Lead Data Scientist at Kanjoya
  • Team: Jade Zhang, Kevin Pollock, Dharshi Devendran, and Shannon Hateley

Our dataset

Every post from Jan-July 2015 (534,305 entries) from the experience project, a social network where users annonymously share life experiences and personal stories.

By asking:

"Which users post to which groups?"

We can learn:

  • if bots / trolls are posting / ruining groups
  • if posting patterns indicate attrition problems
  • if a user can be recommended new groups

Skills

  • Handling and cleaning large datasets
  • Natural Language Processing
  • Supervised and unsupervised learning
  • Python (ipython notebook, scipy, pandas, numpy)
  • R
  • Shiny
  • data visualization

See this repo's ipython notebooks for analysis