CogBrz-Whiting-MC2

VAST Challenge 2019
Mini-Challenge 2

Team Members:

Mark Whiting, Cognitive Breeze, cognitivebreeze@gmail.com

Student Team: No

Tools Used:

Tableau

SensorMap, a mobile sensor tracker built using the NetLogo agent framework

Approximately how many hours were spent working on this submission in total?

<shrug>

May we post your submission in the Visual Analytics Benchmark Repository after VAST Challenge 2019 is complete? Yes

Video

Provide a link to your video. Example:

https://vimeo.com/346760508/98c3596c50

Questions

Your task, as supported by visual analytics that you apply, is to help St. Himark's emergency management team combine data from the government-operated stationary monitors with data from citizen-operated mobile sensors to help them better understand conditions in the city and identify likely locations that will require further monitoring, cleanup, or even evacuation. Will data from citizen scientists clarify the situation or make it more uncertain? Use visual analytics to develop responses to the questions below. Novel visualizations of uncertainty are especially interesting for this mini-challenge.

1 – Visualize radiation measurements over time from both static and mobile sensors to identify areas where radiation over background is detected. Characterize changes over time. Limit your response to 6 images and 500 words.

The video below shows radiation hits by the static and mobile sensors as time passes and the mobile sensors proceed throughout the days of the simulation. The gray lines that appear show paths through the island and can be interpreted as streets. Background radiation is not indicated, but various levels of radiation detected are shown by red, orange, and yellow dots (like breadcrumbs) left on the map over time by the mobile sensors. One of the best ways to use this view is by scrubbing the video control across the simulation days. This view wipes clean the radiation breadcrumbs at 12M each day, so we can more easily see the daily changes. Concentrations of breadcrumbs over time show the areas of potentially most concern over the time period.

Initially on the 6^th, spiked readings can be seen in several of the neighborhoods. Most likely, these are not being motivated by actual radiation events. On the 7^th, there are more sporadic high readings, but again nothing look like a trend.

On the 8^th, there appears to be more hits in Downtown, Northwest, Old Town, Southton, West Parton, and Safe Town.

On the 9^th, there is more activity in Weston, Easton, Southton, and West Parton. The Scenic Vista border is impacted as well. The street to the west of the Easton-SafeTown-East Parton-West Parton four corners is seeing a string of hits.

On the 10^th, the most activity is seen in the Downtown, Southwest, and other sections in the middle of the island, plus on the road leading off the island toward the Jade Bridge. Also, there is some activity on the Wilson Forest Highway heading off of the island .

2 – Use visual analytics to represent and analyze uncertainty in the measurement of radiation across the city.

a. Compare uncertainty of the static sensors to the mobile sensors. What anomalies can you see? Are there sensors that are too uncertain to trust?

b. Which regions of the city have greater uncertainty of radiation measurement? Use visual analytics to explain your rationale.

c. What effects do you see in the sensor readings after the earthquake and other major events? What effect do these events have on uncertainty?

The following view allows you to scrub through the static sensors to observe their behavior over the simulation time. The view shows average values of the sensor readings over time aggregated to the minute level. Each individual sensor remains overlaid across all of the sensors to keep it contextualized within the entire sensor collection.

All of the static sensors are disturbed by an event around 3PM on April 9. Sensor 9’s (Old Town) baseline jumps higher than most. Sensor 11 (Broadview) jumps up and then down from 10PM on the 8^th to 3PM on the 9^th. Sensor 12 (Safetown) has a more exaggerated version of that same effect, with a high offset after the event. Sensor 15 (Safetown, by the reactor) gets stuck from 10PM on the 8^th to 9M on the 10^th.

The next view allows you to scrub through the mobile sensors to observe their behaviors (albeit in non-stationary mode) over the simulation time. Again, these are median values recorded by the mobile sensors aggregated to the minute level throughout the simulation. Again, the sensors are highlighted within the context of all sensors to help compare across the entire group.

The mobile sensors have considerable variance among themselves. Many like Sensors 1 and 2 are fairly flat. Sensor 10 has a big spike on April 8, and 20 has a jump on April 9. Several have long spikes on April 9, including 21, 22, 24, 25, 27, 28, 29, and 45. In the next section, you can see these cars collect at the southeast corner of the island. There are premature ends to 6, 20, 30, 34, 48, and 49. Sensor 12 has a huge spike on April 9, so it is shown at the end.

The following view allows you to see the coverage of the mobile sensors across the island. It also allows you to see the behavior of the mobile sensors over time.

To help interpret this video: hospitals are shown in blue and the nuclear plant is in yellow. The graph to the upper left shows a count of the number of hotspots found with a reading > 1000. Mobile sensors are represented as cars with their sensor number at the bottom right of each. Static sensors are black rectangles with green flair tops. Each static sensor has an identification number of its original number + 5000 to differentiate it from the mobiles. When a hot spot is detected, a “breadcrumb” is left at the location with yellow being lowest level leaving a breadcrumb (>100 reading), orange next highest (>500 reading) and red being hottest (>1000). Mobiles leave traces of their movement to help identify streets across the island. If a car turns white, it has come within a short radius of the nuclear plant and has measured a reading >1000.

Cars 9, 15, and 32 appear to get cooked (sensor 32 doesn’t come within our defined radius, but it stays “pretty close" and is worth keeping an eye on).

Also, from this simulation it appears that the residents are scared off from the Magritte and 12th of July bridges. Or they were blocked by elephants (MC3).

The final view shows all of the locations where high sensor readings were detected and their highest reading over the entire simulation period. There were a lot of hotspots.

The mobile sensors are most present in certain parts of the city including Old Town, Downtown, Easton and Northwest. There is very little coverage for Wilson Forest, Peppermill, and Scenic Vista, among others.

3 – Given the uncertainty you observed in question 2, are the radiation measurements reliable enough to locate areas of concern?

a. Highlight potential locations of contamination, including the locations of contaminated cars. Should St. Himark officials be worried about contaminated cars moving around the city?

The static and dynamic sensor view in Question 2 highlights the potential locations of radiation, both in real time and in aggregate at the end of the video. Always Safe should be worried since they will be liable for health issues and damages resulting from a leak. Particularly with the sewer backup and rain water running through the city, although the rain water may provide help through dilution.

b. Estimate how many cars may have been contaminated when coolant leaked from the Always Safe plant. Use visual analysis of radiation measurements to determine if any have left the area.

Three cars may have been contaminated from the data in this MC. Two are shown turning white in the video (9 and 13). It’s possible that 32 was also irradiated. Movements of the cars is shown, including those leaving the island. (From the information in MC3 however, we’d estimate a whole lot more were cooked during the quake periods).

c. Indicate where you would deploy more sensors to improve radiation monitoring in the city. Would you recommend more static sensors or more mobile sensors or both? Use your visualization of radiation measurement uncertainty to justify your recommendation.

More sensors are needed for Wilson Forest, Pepper Mill, and Terrapin Springs. Gaps in the radiation detection are visible in the final view on the video from question 2.

We would recommend autonomous mobile sensors that cover the entire island in a comprehensive pattern in a reasonable time interval. These should also be re-taskable to focus in potentially contaminated areas when required.

4 –– Summarize the state of radiation measurements at the end of the available period. Use your novel visualizations and analysis approaches to suggest a course of action for the city. Use visual analytics to compare the static sensor network to the mobile sensor network. What are the strengths and weaknesses of each approach? How do they support each other? Limit your response to 6 images and 800 words.

Strengths of static: easier to get accurate baseline, static nature allows for consistency in understanding the situation at a particular location over time.

Weaknesses of static: smaller geographic breadth of coverage

Strengths of mobile: Range of coverage

Weaknesses of mobile: Unknown geospatial destinations, inability to provide long term readings.

If we look at the state of the static and mobile sensors as shown below, we see broad areas of the island uncovered for radiation measurements. Following the pathways of the mobile sensors, you see many of these areas are never covered. Static sensors are sometimes placed in strategic locations, such as sensor 12 (5012 on the map), but miss coverage like on all of the other bridge entry and exit points. Even with the mobile sensors registering a hot spot, there is no guarantee they will return to that spot to continue to characterize it.

5 –– The data for this challenge can be analyzed either as a static collection or as a dynamic stream of data, as it would occur in a real emergency. Describe how you analyzed the data - as a static collection or a stream. How do you think this choice affected your analysis? Limit your response to 200 words and 3 images.

a. This question does not impact scoring: Did you analyze the data as a stream?

Yes and no. The approach using the agent for mobile sensors can be used to look at data as a stream. The post-hoc graphs do not.

b. This question does impact scoring: How do you think your approach (streaming or offline) impacted your conclusions

This is difficult to answer if you had a hybrid approach. It’s more difficult to assess a sensor’s state in a streaming mode unless aberrations are profound.