Understanding the NERSC Community through Network Graphs of Allocation Data

Science/CS domains

Network graphs, social network analysis, information theory

Project description

NERSC is looking for a knowledgeable summer intern to apply social network analysis to our allocations data and report on its potential to identify and respond to community needs and to build a strong NERSC community of practice.

NERSC has an ongoing mission to better understand, serve, and provide for the NERSC user community. NERSC is continuously looking for better ways to investigate current and future needs to better respond to them. As part of this investigation, NERSC has used project allocation data to build network graphs of the NERSC community, including associated code to allow cutting, visualization, and basic analytics.

Example NERSC allocation graph from NERSC’s Jupyter notebooks: Fusion Energy Science allocations, colored by the ratio of their allocated hours used.

The selected summer intern would use and improve upon NERSC-developed tools (written in Python using Jupyter notebooks) to analyze the available data and make recommendations on useful analysis and metrics towards NERSC’s community engagement goals. This is a research-focused project, so some prior knowledge of graph and/or social network analysis is required, and being self-motivated and enthusiastic is a plus.

The outcome of this investigation will be shared in technical papers and community engagement talks and may become the basis for future strategic planning and coordination efforts within NERSC’s community of practice.

Project tasks

Expected tasks for this internship include the following:

Discuss and understand NERSC’s community engagement objectives in order to suggest useful graph analytics information.
Analyze available data, particularly connectivity information among NERSC users, and evaluate its usefulness in describing the community and planning our actions towards our community of practice goals.
Make recommendations on which other data should be collected, analyzed, or tracked to better inform our community engagement goals.
Improve or develop new graph analysis tools, starting from existing Python-based Jupyter notebooks.
Report on the state of the NERSC user community based on the data analyzed and conclusions that can be drawn therein.

Desired skills/background

Experience with Python and JupyterHub
Knowledge of network graph strategies and algorithms
Experience with network graph data structures and libraries (e.g., NetworkX and Plotly)
Knowledge of statistics/statistical analysis and associated libraries (e.g., Numpy and Pandas)
Experience performing scientific research
Experience with literature surveying and algorithm design