Leveraging Big Data for a Better Tomorrow
technical has undergone
a drastic transformation over time
from a regional industrial hub to one of
the
most polluted city in the country to a
city
that now is known to be a living lab for
smart city research
today i want to tell you the story of
how chattanooga
as a mid-sized city in southeast
tennessee
along the tennessee river in the
foothills of appalachian
mountains became a living lab for smart
city
the site was established in 1800s
and the name was changed to chattanooga
in 1834
which means iraq rising to a point it’s
speculated
that refers to the luca mountain around
the city of chattanooga
the city economy was growing
and it became the growth what became
stimulated
when the railroad industry started in
1830s and 1840s
between 18 mid 1800s and mid 1900
chattanooga was a regional hub for
manufacturing
and industry it was producing and
manufacturing
stills saddlery parts automobile parts
and appliances
the city was growing so was the economy
and
population the entrepreneurship
was also there examples of that we can
talk about dixie
and coca-cola bottling company
the manufacturing grew but also it left
scar on the city it was not expected
the rise in manufacturing increased the
pollution
that led to population going down
local economy was not strong enough
and was not strong anymore in fact
the city in 1969 was named the dirtiest
city in the country the city
hit its bottom when it was at its lowest
a group of dedicated people decided
to stay to innovate
to invest they were looking at it
from a high level to see which decision
they can make that it can revolting
the city and
to do that they were looking for
applications that would have
impact on the local
community chattanooga is surrounded by
beautiful natural resources so the
project they chose as the first project
was tennessee river park
this beautiful trail goes along the
tennessee river
those of you that have walked or hiked
or ran on that trail you know that this
is 11 mile
stretch of the trail in downtown area
along the river and you can go on that
without having to
step on any paved road
another example of that was the aquarium
that was built
in downtown area
this group had bold visions
they took bold steps to
completely change the urban they brought
the sense of urban
to downtown area they did a lot of
effort to clean the air to clean the
water
to make urban livable and enjoyable
while innovation was definitely there
technology and entrepreneurship was
still
missing in branding of chattanooga
that was changed when epb installed
fiber optic throughout the city as a
backbone for the telecommunication
that is when epb brought chattanooga to
the age of high speed
to the information age they installed
9 000 miles of fiber that would provide
access to every house and every
businesses
in the 600 square mile territory of epb
that provide internet up to 10 gigabit
per second
at tens of times faster than
national average the gig provided
opportunities for chattanooga to be able
to develop
applications that it would change
completely the way that to do conduct
business
to learn to live and play
and that also provided opportunities
for entrepreneurship to reappear in
chattanooga
geek tank is example of that gig tank is
a
boutique accelerator for startups
to develop business applications that
can thrive
on low latency and fast internet
they bring a cohort of high-tech
companies
here to chattanooga in fact chattanooga
is the only city in the country
that a startup company can test their
ideas
in a community-wide fiber optic
it was around this time that was decided
future is not about technology
future is about people whenever we talk
about smart city
the word smart in smart city resonates
as
sensor internet of things connectivity
data artificial intelligence
that is true it’s not the end of it
the human factor should be behind
every project that it’s done in smart
city
in fact technology should facilitate
application that can
improve the quality of life
but enable in order to be to do that
we need to make sure that community talk
to each other
they talk to people we break the silos
between departments and municipalities
we know what are the challenges and
issues
we need to think about it can research
solve those issues
more importantly can it be implemented
will those solutions be adopted by
citizens
and that is when chattanooga smart
community collaborative
was born in fact that’s the latest
community effort that’s happening in
chattanooga
it consists of seven entities city of
chattanooga
hamilton county epb the enterprise
center
colab which is the startup
incubator the university of tennessee at
chattanooga
erlanger health systems the goal for
the collaborative is to through
collaboration
and through new tools and technologies
we will be able to look at the city
through new lands
there are a lot of efforts going on
through this collaborative
but i’m going to be talking about ones
that are related to
mobility safety and quality of life
about two years ago at utc the center
for urban informatics and progress
cuip was established we are
living in the urbanization century it’s
expected by 2050
which is only 30 years from now
two-thirds of population living in urban
environment
it has its own challenges and its own
opportunities
the goal at cuip is to address these
challenges
we want to make urban livable accessible
healthy for all at cuip
we do applied researches research that
would contribute to the smart city
research body as well as improve the
community of life
it would contribute to the local
community
and we are working on different areas
health energy mobility social science
and we’re going to talk about couple of
them today
which city doesn’t want to reduce their
number of crashes
in fact numbers are showing that the
number of accidents
in tennessee and in hamilton county as
well as the number of fatalities
have been growing over the past few
years
9-1-1 data is the open source data that
everyone can have
access to it we wanted to see that what
does these numbers tell us
what can we do with these numbers and
how can we use them to remedy these
issues
we drew them with respect to time we
aggregate data
and we looked at them based on the time
of the day
for every day of the week the trend
clearly shows the number of accident
goes up during the rush hour
and the number of accidents during the
weekday is more than the weekends
in fact it even shows us the rush hour
on fridays it starts at 4 pm
the other thing we wanted to look at was
how about spatially so we look at the
data according to space
and as you can see here this shows the
number of
aggregated accidents that happen at each
space
changing over the time clearly there is
a trend there
at the same time we talked about this
that perhaps everybody
knows accidents are happening on the
where accidents are happening on
interstates
and there are more accidents there and
also everybody probably knows
where are these accidents happening
where are the exact locations on the
interstates
so we wanted to look at more of a local
intersections
that they were having a high number of
accidents there were hot spots for
accidents and crashes
what you can see here is a residential
street
that has a speed limit of 35.
in the larger region there has been 88
accidents in the past two years and in
that exact
spot there has been 22 accidents
this is a objective and unbiased data
while we are data scientists we’re not
necessarily traffic engineers or urban
planners
so true collaborative we reached out to
the sparsity director of city of
chattanooga
through the discussions with him we
realized that
the poll was the reason for the car
accidents
the pool was too close to the road by
moving it ever so
slightly there has been no accidents
since june 2019.
this is a project that has a real impact
on community
we are saving lives we are saving
resources
saving property and also we’re saving
human cost here
but these are the things that we can
learn based on
historical data the question
is can we do prediction can we predict
where are going to be the accidents on a
thursday rainy afternoon
and that’s what we decided to do we
looked at the data
the 911 data is openly available
we also looked at road geometries how is
the curvature of the road
how many lanes there are there are there
sidewalks there
are there crosswalks there what kind of
pavement it’s there
and we also look at the weather at the
time of accident
was it raining was it foggy
or was it a clear sunny day we put all
these data together and we built a
predictive model
what you see here is a prediction that
we have done
on the accidents for january 20th 2020.
we use the data up to january 19
to predict where are going to be the
accidents
on january 20th and what you see the
hexagons
show the areas that have the chance of
accidents
the bluer the color the higher rate
accident the higher chance of accident
happening there
on january 21st we got the data from
9-1-1 county and we looked and we
overlaid them
and you can see that we can predict with
the high accuracy of where the accidents
potentially can happen one of the
immediate applications for this is we
can
allocate first responders according to
this prediction
we all know that the faster the response
the lower the lower the fatality rates
and also property damage
so we talked about vehicles what about
pedestrians
what about cyclists even if you drive to
your office you still have to park and
perhaps cross the street to get to your
office
or you go for a cup of coffee or go for
lunch you still have to walk or bike
we wanted to also see how it or
how safe are our roads where are the hot
spots of
accidents for pedestrian and cyclists
because we have seen the
fatality rates that were related
to car crashes for pedestrian are also
rising
to be able to do that we need to answer
some questions
how and when a pedestrian would use
a given sidewalk
how how are they going to be using
different crosswalks
how long do they have to wait to be able
to cross that intersection
because if you have to wait a long time
to cross perhaps you’re going to jaywalk
the same thing for vehicles how long do
they have to wait to before they can
cross the intersection
because if it’s going to be too long
then they’re going to be more aggressive
the answering to these questions would
give us actionable data
that we can look into to evaluate the
safety for pedestrians and cyclists
but unlike the 911 data this data
doesn’t exist so true collaborative
we actually built a test bed in downtown
chattanooga
those of you that are from chattanooga
recognize this history this is martin
luther king boulevard in downtown
we have built a test bed which is a mile
and a quarter route stretch of the
boulevard here and we chose this
because it has a bike lane it has bike
share station
ev charging station it has a transit
it’s next to campus there’s a lot of
walking and biking activities happening
there
also has residential and businesses so
it kind of represents a small version of
the urban
we equipped different intersections with
different sensors
such as cameras we have high resolution
cameras that we can detect the objects
what we care about is this is a car a
person
a bus we do not care about the
personally identifiable data there
we also have lidar there that would be
able to give us the distance in a high
resolution
we have audio sensor perhaps we can use
it to hear
siren and be able to empty the street so
emergency vehicle can pass through
faster
there are also 5g and fiber
for being able to transmit data with low
latency
and finally we have also edge computing
for applications that
require processing fast and right on the
spot
what you see here is one of the cameras
on the test bed we have been collecting
the data we have been using computer
vision
to detect the objects and after
collecting them over
a period of time what you see at the
bottom here is a hit map
of the objects that have been seen on
the test bed over a period of time
pink shows where the pedestrian are
it makes a perfect sense that on the
sidewalk is we see a lot of pink
also next to the parking spots we see
pink because people
can park their car and get out of their
car but we also see
there are some jaywalking activity there
are some it’s been seeing that
there’s some crossing happening at
non-designated areas
we also see that there are places that
the cars and buses are getting
too close to the bike lane using this we
will be able to evaluate the safety of
our streets
similar to the previous case these are
history data
what about prediction can we do any
prediction here
what you see in this video shows us
where is every object going to be in the
next
couple of seconds so if i know the
person that is crossing where it’s going
to be in the next couple of seconds
and where the vehicle going to be in the
next couple of seconds then we can
measure the time to collision if that is
too small
that should be a warning for the driver
as well as the pedestrian or the cyclist
using this
we can collect where are the places in
the city
that we are seeing a lot of near misses
which basically means the accident
didn’t happen but it was about to happen
and knowing those areas can improve the
safety of
our streets
so in summary this happened in
chattanooga
because a dedicated group of people
decided
to stay invest and innovate
same choices are happening today
we can see that the innovation and
investment that happened in the
collaborative
in cuip the testbed that we just talked
about
even in the past few years we have seen
how smart city had positive
outcome in our community
the collaborative is involved in a lot
of other projects in addition to what i
talked about related to mobility
they’re doing projects in energy in
health equity
social science and more
these efforts that are ongoing will take
chattanooga to the next level
perhaps some of these can also be
used in other communities they can be
adopted by other communities as well
thank you