ISSS608 2016 17T1 Group6 Project Report - Discussion
| Proposal |
Contents
Discussion
- Key point, to evaluate the system that we have built*
Feedback from audience
New insights/ practices from the system
Future Work
Visualisation Recommendations
We wish to improve our visualisation dashboard by linking the Text Cloud to the cross filter. At the moment this is not achievable as we could not keep the time dimension after breaking the words into tokens.
We also hope to add tooltips to show the terrorist attack details when the mouse hovering on different parts of the map. We give up on showing the details of individual events because there will be too many points on the map, which makes the web page extremely slow and prone to crash. However, we believe there should be a better way of adding event specific information.
Moreover, we wish to improve our dashboard to allow sorting on total number of attacks, targets of attacks and types of attacks.
Dataset Recommendations
The dataset has too many irrelevant words and missing data in the motivation column. Most of the motivations records tend to be about violate incidents in the United States, whereas the attacks that caused large numbers of casualties often have no motivation records. Clearer motivations will help identify the keywords in the Word Cloud and come up with strategies to tackle the root cause of terrorist attacks.
The dataset also has a large proportion of unknown information or missing records, which makes the analysis of certain aspects difficult. For instance, the classification of International Attacks has around 60% of the records being unknown. As a result, we could not use the variable for analysis as it does not show a clear enough picture. The data needs more updates to be more complete.
