Difference between revisions of "Project Proposal"

From Visual Analytics and Applications
Jump to navigation Jump to search
Line 68: Line 68:
  
 
===== Making the edges table =====
 
===== Making the edges table =====
 +
 +
 +
=== Developing a network graph using ‘ggraph’: ===
 +
 +
 +
 +
=== Developing a network graph using ‘ggnetwork’: ===
 +
 +
 +
 +
 +
=== Adding interactivity to the plots using ggplotly: ===
 +
 +
 +
 +
=== Making the shiny dashboard: ===
 +
 +
Design Framework:  A detail description of the design principles used and data visualisation elements built (Refer to Section 3: Interface of this paper [1].
  
 
==Assumptions==
 
==Assumptions==

Revision as of 10:26, 5 August 2017

Discovering traffic patterns by using network graph visualisations

Introduction

About

Project Proposal

Project Timeline

App & Deliverables

Poster

 

Project Proposal

Steps in planning and preparation of the application

  • Making sense of the data provided.
  • Selecting a real world practical use-case (Traffic networks).
  • Choosing R and deriving maximum value of the ggraph and ggnetwork packages.
  • Making the dataset reactive to user inputs and choosing the appropriate inputs.
  • Compiling the deliverables to make a complete story through an interactive application.
  • Drafting poster for quick readability and preliminary understanding.

Using R to visualise traffic networks

A critique on previous works and why R might be more suitable?

What are the R packages needed?

User Guide to develop the visualisations

Data transformation

In order to make the network graph render, a well defined nodes and edges table is necessary. Once the main data table is uploaded onto R, the data manipulation functions from R are used in order to achieve the objectives to segregate into a nodes table and edges table.

Deriving new variables for user interactivity:
Making the nodes table
Making the edges table

Developing a network graph using ‘ggraph’:

Developing a network graph using ‘ggnetwork’:

Adding interactivity to the plots using ggplotly:

Making the shiny dashboard:

Design Framework: A detail description of the design principles used and data visualisation elements built (Refer to Section 3: Interface of this paper [1].

Assumptions

The main underlying assumption lies with the mapping of the travel route for each vehicle. The data does not provide the GPS location of the cars at the different timestamps. Hence, a sorting of the timestamps is done based on records provided by the different entrances and it follows that a car travels directly to the next entrance after passing the current one where in fact it could have made a detour or take other routes without passing any entrances hence no data is being recorded.


Limitations

The ggraph package needs a well defined nodes and edges table in order to produce visualisations. While R Shiny enables development of quick and open source applications, extensive data transformation and reshaping is needed from the dataset in order to make full utilisation of the package for seamless performance of the application.

Future Scope

Future Work - A description of how your system could be extended or refined.


With the help of the timestamp and coordinate information of specified nodes, speeds of various vehicles can be derived, since the distance travelled and time spent between any two nodes are known. This will help to understand corridors in a vicinity where most speeding incidents occur, where there is higher congestion, etc. Also, at corridors with higher congestion typically in rush hours or after work hours, ERP pricings can be revised to divert the traffic to less congested areas.