Navigating Information Landscapes with D3-based Web Apps

Authors: Chris Clasen, National Geospatial Intelligence Agency, Eliza Bradley*, Pennsylvania State University, Robert Fraleigh, Pennsylvania State University, Shawn Hough, Pennsylvania State University, Amber Davis, Pennsylvania State University, Patrick Yuen, National Geospatial Intelligence Agency, Kateri Garcia, National Geospatial Intelligence Agency
Topics: Geographic Information Science and Systems, Remote Sensing, Temporal GIS
Keywords: web, apps, scientometrics, visualization
Session Type: Poster
Day: 4/5/2019
Start / End Time: 8:00 AM / 9:40 AM
Room: Lincoln 2, Marriott, Exhibition Level
Presentation File: No File Uploaded


Rich information landscapes (academic publication records, machine learning detection results from remote sensing data) can provide tremendous insight, but are often difficult to navigate and summarize in meaningful ways. Our work highlights using JavaScript libraries (D3.js and DC.js) for two prototypes on different datasets: (1) LitExplorer - for exploring the landscape of LiDAR academic research through space and time, and (2) D3tector - an interactive framework for visualizing object detection results in attribute space with direct linkage to an image chip display. LitExplorer offers a user the ability to crossfilter keywords, institutions, authors, and additional literature metadata associated with queries from Microsoft Academic’s Cognitive Services API. The query results are displayed on a map, table, and in summary form (histograms based on different attributes). D3tector provides a way to visualize data not just on a map or in table form, but in attribute space. This is relevant for analyzing object detection results (e.g. object size, dominant color, nearest distance to another object). As a user moves the cursor over data points in D3tector, it updates the display to show the corresponding imagery chip and other corresponding graphics. This accelerates evaluations for distributed detections over large areas, highlights trends in false positives, and reveals significant context (e.g. co-located secondary objects or atypical arrangement/orientations).

Abstract Information

This abstract is already part of a session. View the session here.

To access contact information login