Entry Name: "SMU-Team
JKY-GC"
VAST Challenge 2017
Grand
Challenge
Team Members:
Dr. Kam Tin Seong, Singapore Management University, tskam@smu.edu.sg
Kishan Bharadwaj Shridhar, Singapore Management University, kishanbs.2016@mitb.smu.edu.sg
Ong Guan Jie Jason, Singapore
Management University, jason.ong.2016@mitb.smu.edu.sg
Zhang Yanrong, Singapore Management
University, yrzhang.2016@mitb.smu.edu.sg , PRIMARY
Student Team: YES
Tools Used:
- Tableau
- SAS
JMP
- MS
Visio
- QGIS
Approximately how many hours were spent working on this submission
in total?
180
May we post your submission in the Visual Analytics Benchmark
Repository after VAST Challenge 2017 is complete? YES
Video
Tableau Workbook
Questions
1 Provide your best hypothesis with
supporting evidence of what is happening in the Lekagul Preserve that is
affecting the Rose-crested Blue Pipit. Your answer should include identification
of who is responsible for the impactful activities, what they are doing, where
impactful events are occurring, when and how often these occur, how these
events are taking place, and why they are happening.
Please limit your response to 1000 words and 6 images.
Provide your answer and corresponding images here.
The links between traffic patterns
in the preserve, the image degradation as shown by satellite imagery and the
emissions of gases by factories have a combined effect in affecting the rose-crested
blue pipit.
An overview of the events happening
across the preserve and the industrial park is as shown in the spider web
diagram below. The key information gathered from mini challenge 1, i.e.
visitors to various parts of the preserve, and the anomalous behaviour of
certain vehicles is traced to the emissions of gases by factories. To further
validate the findings, the satellite imagery is used as a supporting evidence
to link the locations of the events to the timeframes they occur.
Hypothesis #1: 4 axle trucks in the preserve are helping out
Kasios in dumping Methylosmolene, which leads to degradation of plant health in
the Ranger stop-3 region. This is having an adverse effect on the birds, as the
chemicals are harmful to vertebrates.
- Evidence
#1: 23 4-axle trucks trespass to travel in Gates. The pattern of paths
shows that they use the same predefined path from Entrance 3 to Ranger
Stop 3, and then back. This is a repeating pattern, happening from 2am to
5am on a given day, for all months from May 2015- May 2016, except for
April 2016.
The image shows the predefined
paths in which the trucks are travelling, shuttling back and forth between
Entrance 3 and Ranger Stop 3. The anomaly arises out of the time in which they
are found, and in the fact that they are traversing through Gates, which are
supposed to be inaccessible for non-ranger vehicles. 23 different trucks are
following this pattern. Measuring the speeds of these trucks further
illustrates the average speed is always maintained higher than 25mph, meaning
they might be in a rush through the corridors they pass by, disturbing the
sleeping rhythms of the blue pipit.
- Evidence
#2: Kasios releases Methylosmolene from 10pm to 6 am on certain days in
April 2016, despite previously complying to environmental norms in June
2013
The cycle plot above shows that methylosmolene is typically
emitted during the wee hours of 22:00 - 06:00.
The air plume model on the other
hand helps to associate the release of Methylosmolene to the Kasios factory. In
this snippet taken on 25th April 2016 at 03:00 am, it is clearly evident that
the polygon representing the stream of wind detected by monitor 6 contains
Kasios in its trajectory. Hence, it is highly likely the company is directly
responsible for emitting that chemical.
- Evidence
#3: The Ranger Stop 3 experiences a degradation of plant health
As seen from the GIF and the Air
Plume model above, there is a correlation which can be identified with the 4-axle
truck, the Kasios emissions and the satellite imagery showing degradation of
plant health. The degradation of plant health and other key evidence from the
satellite imagery is as shown below. The ranger stop 3 point is
annotated, and over a period of two year gap, between Sep 2014 and 2016, it is
seen that there are degradations of plant life around it.
This hypothesis is further
supported by the information obtained from the Mistford newsletters. Kasios is
looking to hire more transport personnel and has an employee testimonial,
stating that he was rewarded for working during odd hours. This gels well with
the fact that the times of operation of those trucks were 2am to 5am, which is
when the Ranger vehicles within the preserve are also not in operation. The
timeline below shows how they all occur coherently.
Hypothesis #2: The sites of the preserve, where people
throng such as Camping 5, experience increased degradation in plant life due to
increased footfall.
- Evidence
#1: Camping 5 is the site with most number of traffic, but overnight
campers are found only in the months of July and Aug 2015.
It is known from the analysis of
traffic that camping 4 and camping 5 are the camp sites with the most traffic
across the observation period. It is also clear from the satellite imagery
showing that camping 5 also experiences degradation in plant health over the
years. An aggregate of over 115,000 hours were spent by all vehicles at camping
5 alone. While day campers do not cause camp fires, it can be said that the
people who tend to stay within camping sites, from Late Evening to Early
Morning (i.e. from 6pm to 6am on a given day), might be causing disturbances to
the bird habitats.
- Evidence
#2: Boonsong lake is located near camping 5.
From the identification of the
Boonsong lake near camping 5 in the satellite image above, it can be inferred
that the scenic nature of camping 5 is more appealing, which is the reason it
is drawing higher crowds into the preserve.
2 Provide a timeline that comprehensively
describes the relevant activities in Mistford, the Industrial Park, and the
Preserve that helps concisely describe the events identified in Question 1.
Please limit your response to 1000 words and 6 images.
Time Stamp
Key Events
|
June 2010 |
Mistford Environmental Authority
set stricter emissions standards for volatile organic compounds (VOCs). |
|
June 2013 |
New emissions standards for
volatile organic compounds (VOCs) take effect. |
|
June 2013 |
Kasios’ stock price begins to drop
sharply. |
|
Early 2014 |
Kasios' three-year expansion
project which will increase the furniture manufacturer's capacity begins. |
|
Jun 2014 |
Kasios’ introduces its new line of
executive office chairs called Excalibur. |
|
Jul - Sep
2014 |
Kasios’ recruits more than thirty
process engineers, mechanics and transportation experts over the quarter. |
|
Dec 2014 |
Kasios’ stock price begins to
recover. |
|
Mar 2015 |
Kasios’ stock price begins to surge
higher. |
|
May 2015 |
Kasiosstyle.com, a new ecommerce
platform is released by Kasios. |
|
Jun 2015 |
Robert Steuhaven is named to board
of directors of Kasios. |
|
Jun
2015-Aug 2015 |
Surge in the traffic across the
preserve. |
|
Sep 2015 |
Traffic entering the preserve
begins to decrease sharply. |
|
Dec 2015 |
John Torch, Kasios transport
worker, mentions that hours are sometimes long and hard on the road he has to
work pretty odd hours sometimes. |
|
May
2015-May 2016 |
4-Axle trucks are found to loiter
suspiciously at higher speeds between Entrance 3 and Ranger Stop 3 and back
again, from 2am-5am. |
|
Apr – Dec
2016 |
Detection of the release of
Methylosmolene typically from 22:00 – 06:00. |
|
Sep 2016 |
Satellite image shows the health
condition of plants have turned worse around Ranger Stop3, Ranger Stop1 and
Ranger Stop2 compared to last year. |
|
Aug – Dec
2016 |
Higher average monthly readings of
Methylosmolene detected on almost all the monitors. |
|
Dec 2016 |
Satellite image shows snow coverage
has decreased and the moisture becomes higher as compared to last year. |
Kasios is one of the factories
which is under the control of the new emissions standards for volatile organic
compounds (VOCs). But right after the standards take effect in June 2013,
the stock price begins to plummet because they are forced to source other
high-cost alternatives for Methylosmolene. However, a recovery point occurs
across the quarter of Jul to Sep 2014 when they introduce a new line of
executive office chairs called Excaliburr.
A collection of hints which are
collected from the newsletters are as shown above. Clockwise from top left,
Environmental compliance of Kasios, Kasios’ statement on expanding workforce,
John Torch’s testimonial, Kasios’ expansion plans.
Kasios recruits new transportation
experts and process engineers during the same quarter. Since no data on sensors
is available for 2014 and 2015 for sensor data, some assumptions are made that
sensor data being captured in 2016, were applicable in the previous years.
A trend is seen from the sensor
readings, where Methylosmolene is being detected across the three months,
namely April, August and December 2016, from 10pm to 6am. The assumption made
here is that Kasios had decided to reuse Methylosmolene as part of their
process, and needed a way to dispose off the effluents through chemical
neutralisation, which is a requirement of the chemical. This gels well with the
fact that they needed new process engineers, and we can assume that they have
releasing Methylosmolene right through 2014, 2015 and the detected periods in
2016. The sensor readings availability in 2016 can be backtracked and the plot
can be stitched together.
Also, it is seen the anomalous 4
axle trucks looping between Entrance 3 and ranger stop 3, trespassing across
gates, and moving at speeds which are above the limit, from 2am to 5am., from
May 2015 to May 2016. Linking the events can illustrate that gates, which are
not supposed to be accessed by non-ranger vehicles, are being used as a pathway
by drivers who work for Kasios. This is also coincident with an employee, John
Torch who is getting rewarded for working odd hours in December 2015. Kasios
drivers might be doing this to avoid detection by ranger vehicles, which are
generally found operating in the time range of 6am to 10pm on a given day. The
coherence is not only established by the overlapping timelines but also due to
the proximity of Entrance 3 to the region where Mistford industrial park is
located.
From Mar 2015 onward, it is also
seen that Kasios begins to dramatically boost their bottom line, which might be
due to three reasons.
- It is illegally producing Methylosmolene,
which is helping it to cut down costs.
- The appointment of Robert Steuhauven and
his diversification strategy is aiding recovery, or
- Their new e-commerce website is garnering
traction.
The claims above are further
supported by satellite imagery obtained in Sep 2016, which shows that the
region around ranger stop 3 has experienced deterioration in plant quality as
compared to Sep 2014. Also, temperatures in the preserve have dropped when
compared to the two winters of Dec 2016 and 2014, as illustrated below. The
histogram shows the shift between the years, in terms of the snow cover as well
as the moisture content in the area.
Tying this to the traffic patterns show that this may explain why
lesser traffic is being found in the preserve post July 2015, which is when
summer typically ends.
Kasios’ and few other factories in
the preserve have plans to expand on their facilities, one evidence of which is
found from newsletter released in Jun 2013. However, since the satellite
imagery does not cover the Mistford industrial park, further inferences are
being hindered.
3 How confident are you in your
hypothesis? What factors impact your confidence in your hypothesis? What
additional information would help strengthen your hypothesis? Please
limit your response to 500 words and 3 images.
Provide your answer and corresponding images here.
Timeframes of available data:
The period of data availability for
the traffic patterns is incoherent with the time frame when data from the
sensors is available. Also, the practical impossibility of obtaining satellite
images for continuous time intervals, introduces forced assumptions. For e.g.
patterns observed in June 2014 in satellite imagery is typically assumed to
exist until the next available image.
Testimonials lend higher
credibility:
The testimonial from Kristina, HR
personnel from Kasios, showing that they began to hire transportation personnel
along with the testimonial of John Torch, the employee of Kasios who was
rewarded for working long hours in December 2015, add key weights in boosting
the evidences collected with regard to Hypotheses #1. More credibility can be
lent into the hypotheses if there is information on what John meant by working
odd hours, and what those odd hours were.
Instances where more data would
have lent clarity:
Traffic data of the preserve beyond
Entrance 3 and Entrance 4, would have been helpful to validate if the 23
vehicles in Hypothesis #1 indeed travel from the roads starting from Kasios.
To estimate the air plume model, an
angle of 10 degree is assumed, but if the exact information on the aperture
that the sensor is able to capture is available, it can help to visualise more
clearly which factory is causing the respective emission.
So, while the changes in temperature profile are available,
missing information regarding the blue pipit habitat’s climatic preferences
render it unusable to the full extent.
The stock price of Kasios begins to
increase from May 2015. This might be due to the support of the nefarious
activity which is indirectly boosting their sales bottom line, i.e. they are
dumping the unwanted chemicals in the preserve, in order to boost their
corporate image, and that they indeed were compliant to environmental
regulations. However, it can also be due to the launch of their new E-commerce
website. Another factor could be the successful diversification of company
strategy implemented by Robert Steuhaven, who was appointed to the board
of directors. The mix of the factors lends fuzziness to concluding if stock
price and Hypotheses #1 had a clear correlation.
Evidence from newsletters point to boonsong lake is used for
fishing activities, which might be the reason, campers might want to stay
longer, to get some fish. Also, summer months might be coincident with the fish
breeding season, due to which there is increased footfall in those camping
sights, but without exact information, the hypotheses cannot be fully
constructed, on why there is preference for certain camps. This might show why
the balance of the footfall in the camps are unevenly distributed, for e.g.
sites camping 0 & 1 have the least footfall in the entire observation
period.
Clouded satellite imagery:
Satellite images tend to be filled with clouds as an example is
shown below during November 28, 2014. This impeded the observations that could
be carried out, in those periods.
4 What are your proposed next steps? Do
you have a course of action to correct the problems in the Lekagul Preserve and
help the Rose-crested Blue Pipit? Please limit your response
to 500 words and 3 images.
Provide your answer and corresponding images here.
Step #1: Measuring the blue pipit
population and their habitats
It is assumed that the preserve
management do not have quantifiable figures on the estimates. Having
quantifiable estimates will help future analysts to see the rate of decline and
correlate to the patterns they observe. A way of estimating the bird population
could be the traditional mark and recapture technique, which might yield
initial estimates. With latest advances in deep learning and image recognition,
cameras can be installed at key places in the preserve borders to estimate
better.
The underlying logic can be
illustrated with the below image, where trained images of the blue pipit can be
tagged using deep learning networks, and the counter of predictions where it
identifies the blue pipit can give a better estimate of how much blue pipits
were found in a zone.
Image Courtesy: http://code.flickr.net/2014/10/20/introducing-flickr-park-or-bird/
This can help the management to
identify blue pipit nesting hotspots and regulate traffic flux through a quota
system across those corridors. Specifically, night time access might be
regulated and stricter laws such as designating certain zones within campsites
for bonfires, can be implemented, to prevent frequent disturbances due to light
and sound people might make when gathering in large numbers.
Vehicles are found to exceed their
speeding limits during the time range of midnight to morning 6am, and such
incidents can be solved by using sensors which readily transmit speed data to
the preserve office, to prevent frequent incidents of overspeeding, which might
cause noise disturbances or damage to ecosystem.
Step #2: Obtain satellite imagery
for the Mistford industrial park
The image above shows satellite
coverage in green hue is outside the region where Mistford park is located. If
available, the data can help analyse if the expansion of factories and
emissions through the air pollute the clouds and have a combined effect in
affecting the blue pipit. Also, the Lekagul management can conduct audits on
factories, especially Kasios, to check on their environmental compliance
satisfactions, and work collaboratively with environment experts to find out if
the type of chemicals used are indeed affecting the blue pipit population.
Step #3 Regulation of Factories within the Vicinity
It is very puzzling when a natural
preserve is placed at a location where there are factories in the vicinity or
why are factories being allowed to be built when there is a natural preserve
nearby. Furthermore, two of the four factories are even undergoing further
expansion which has already been approved. This will only serve to exacerbate
the current pollution problem. Ideally, if this area is designated for the
natural preserve, factories should be made to relocate elsewhere. However, this
is easier said than done because the existence of the natural preserve may be
attributed to the funding by factories.
Alternatively, regulations can be
passed to have monitors installed directly at the four factories themselves to
ensure their conformity with regards to the emission of prohibited chemicals.