Projects Underway

Developing an Interactive Web-Application for Instructions Involving Networks

Changhyun Kwon

Project Team
  • Ann Bisantz, University at Buffalo
  • Joan Lucas, SUNY College at Brockport
  • June Dong, SUNY Oswego
  • Julia Colyar, University at Buffalo

Buffalo, University at

2013

Tier Two

Project Abstract:

1 Background and Objective

The objective of this Tier-2 project is to use the most recent web technology to improve educational effectiveness for teaching scientific concepts and computational methods in a connected world. This project will develop an interactive web-application that will be readily sharable with all SUNY instructors who teach networks. Nowadays, the concept of networks is a common topic taught in many disciplines: biology, business, communication, computer science, transportation engineering, industrial engineering, sociology, and so on. A network may be the Internet, road network, food chain, supply chain, wireless network, or social network. Difficult concepts that are often written in complicated mathematical notations weaken the teaching effectiveness of the instructors and the learning efficiency of the students. Consequently, developing an environment to easily interact and share scientific notions is required. This project is clearly Tier-2, as it develops a new innovative application.

A network consists of a set of nodes, a set of links, and a topology that describes how nodes and links are connected. Each node and each link has its own properties. For example, in a road network, a link is a road segment connecting two intersections, and it can have properties like distance, travel time, congestion level, accident probability, population around it, number of lanes, speed limit, etc. A path is an ordered collection of links to connect one node to another. In a social network, each node represents a person, and a path can mean how I can connect with an arbitrary person through my friends of friends.

Often, we measure characteristics of a path, and determine the best path among many possibilities. The most widely used example may be finding the shortest path, which is easily found in GPS navigation systems for vehicular networks. In a social network, one can measure the number of friends of friends required to connect with a person to quantify connectivity, and in a telecommunication network, one can measure the expected time for a signal to reach its destination and the probability of losing the signal. Based on such measures, we can determine the best path by solving a mathematical optimization problem.

Teaching how we measure and explaining why a path is better than another can sometimes be difficult. For a shortest-path finding example, if the network is drawn to reflect the actual length of links and the actual geographical location of nodes, then humans can easily find a reasonably short path fairly quickly. It is because the visualized network displays useful information. However, if each node and link has multiple properties with probabilistic behavior, displaying information in traditional pedagogical methods is difficult and computing a measure may take a long time, which will delay learning experiences of students. This project aims to develop a web-application to help instructors to better deliver concepts and methods in networks and students to better understand those by playing with the web-application.

2 Sketch of the Web-Application

This section first describes what the web-application will look like in hazardous materials transportation networks, which is the main teaching and research interest of the principal investigator (PI). Please note that the proposed web-application will be developed for general network-related instructions. A schematic idea of the web system screen is sketched in Figure 1 (Find in Reports and Resources below). On the left part of the screen, a user sees a network and selects links to make a path from an origin to a destination. For each link, the user can see the attribute values like accident probability and accident consequence. As links are chosen and a path is formed, on the right of the screen the user can see the measures for various path performance criteria. Users will use client modules on a web-browser and an administrator will have options to generate random hazardous materials incidents for evaluating the selected path in the current simulation. In Figure 1, TR, MM, VaR, CVaR, and SRM are various risk measures, some of which require complicated computations. The web-application will automatically compute them and allow the user to quickly evaluate the selected path.

Although only textual information is sketched in Figure 1, the developed web-application will additionally use visualization techniques to display information. For example, we may use width of links for accident consequence, and colors of links for accident probabilities. The Co-PI, Dr. Ann Bisantz, is an expert in human-computer interactions and information visualization, and she will suggest the best way to display network properties depending on the context.

The web-application system will receive inputs from users (i.e. instructors) to accommodate their different needs in various disciplines. We will first identify the types of networks that are most common in the current SUNY course offerings, and after classification, we will define a pre-set information display scheme for each type. Then, instructors will upload their own network data: nodes, links, topology, and properties.

Since the proposed web-application is interactive, an instructor can easily compare paths with various properties, use visualization and simulation tools and respond to students’ questions such as, “I think that path is better. Can we try it?” In addition, in a computer lab, students can use the web-application to test their ideas and the concepts learnt in the class by themselves. Even further, the web-application system can be used for a competition. (See Assessment Plan.)

3 Plan for Development and Maintenance

Since the PI had worked as a professional web application developer for three years, his experience strongly supports the project in terms of how a web application would be developed and maintained; he will also contribute to the actual development of the system. For the development, D3.js, jQuery, PHP, and Python will be used. D3.js and jQuery are JavaScript libraries for data visualization (HTML5, SVG) and client-side HTML scripting (Ajax, CSS), respectively, while PHP is a server-side scripting language and Python is a server-side programming language suitable for scientific computation. Two undergraduate students in Industrial Engineering or Computer Science will be involved to help development and maintain the system. Please note that JavaScript is a totally different language from JAVA that has been recently criticized for its security issues.

Project Outcome:

Outcomes TBD