ISSS608 2016-17 T3 Assign ZHOU YUHUI Conclusion

From Visual Analytics and Applications
Jump to navigation Jump to search
Qbird.jpg


VAST Challenge 2017:Mystery at the Wildlife Preserve

Background

Methodology & Data Preparation

Insights

Conclusion

 


Question4

The top 3 suspicious patterns, as far as I am concerned, are:

1. No.of Stops vs Overall time spent

Those 3 Car-IDs that stayed in the park for extremely long are very suspicious. For other cars, their car-ID is generated whenever then enter the park at entrance gate. Yet these 3 Car-IDs have more than 2 records of entrance gates, which means their car-IDs are not per-trip generated. This is not reasonable.

2. "Gate" Visitors at night

Gate is gate type for rangers only. Any non-ranger car that accessed this gate type is considered suspicious. Especially when these records happen at late night, it is even worse.

3. "ranger-stop" Visitors at night

Ranger-stop is also for rangers only. Same reason as why Gate visitors at night are suspicious.


Comments

Hi Yuhui,

Well done! You've provided very clear analysis and explanations. I liked the explanations on the rationales for choosing different types of visualizations in addressing different problems especially, a clear explanation on the methodologies would allow the readers in understanding the entire analysis better. You also provided an interesting analysis on the summer camps in the section for the longer period patterns, this is something I've overlooked in my analysis.

For suggestions on the future improvements, since you have already created the preserve map and extracted the coordinates of different gates, maybe you could link it with some of the analysis (eg visualize on a map where did the suspicious cars go) to give the user a better picture on what is going on in the preserve.

Yifei

Bixin.png

Hi Yuhui,

First of all, great job on your assignment. All questions have been covered and clearly illustrated in your work.

Referring to your dashboard, the workbook is readable and understandable.

For suggestions, overall it's already a really good analysis work, and if time permitting, I believe you would definitely explore more forms of visualization as what you usually do.

All the best.

Cao Bo