Visualizing Graduate Admissions Data

Final project for CMPS 161: Introduction to Data Visualization, Winter 2017.

Abstract: The unusually large applicant pool to the Computer Science graduate program at UCSC in 2017 introduced a problem of scale to the admissions process. By applying multivariate visualization techniques to applicants' admission data, I aim to identify correlations in the data that could be applicable to the selection process in the future. I employ a common information visualization technique known as the parallel coordinate plot. A brief overview of this technique is provided, as well as descriptions of several common methods to improve its readability and usefulness that are implemented in this project.

This project was implemented with d3.js 4.7.3, using modules d3-axis, d3-scale, d3-selection, d3-brush, and d3-shape.

Implementation: demo.html
Technical Documentation: Writeup.pdf
User Documentation: README
Source Code: main.js, process.py
Datasets: cleaned.f17.csv (Original), cleaned_new.csv (Processed), cleaned_reduced.csv (Size Reduced)

To use the Parallel Coordinate Plotter, open a .csv file using the file input, and click the load button. The program will parse the input, and display a parallel coordinate plot of the data. By default, only the first 5 fields will be mapped to axes and displayed in the output, and the color mapping is "None".

Expand any of the control panels on the left by clicking on them. In the "Display Axis" panel, check the boxes to choose which axes to display in the plot. In the "Groups" panel, choose which grouping to use to map color to the data. In the "Legend" panel, see which colors are mapped to which sets, and choose which sets to highlight.

The plot itself currently supports interaction only in the form of brushing. Click and drag on an axis to select only records that fall within that range on the axis.

Sample datasets are included in the "datasets" folder.