►
Description
September 2020
Presenter: Dr. Aryya Gangopadhyay
Institution: UMBC
Title: "Data Science for Social Good in Urban Areas"
View Speaker Notes - https://drive.google.com/drive/folders/17wWOdk9MKUMJDNJK4Lv6evcTnq5CkXov?usp=sharing
http://sbdh-prod.ideas.gatech.edu/resources/newsblog/education-and-workforce-working-group
A
So
this
is
the
poster
that
we
presented
in
the
april
pi
meeting
and
it's
a
collaborative
proposal.
So
there
are,
there
are
three
four
institutions.
Actually
umbc
is
the
lead
and
then
we've
got
towson
university
and
we
have
a
bowie
state
university
and
we
have
got
the
university
of
baltimore.
So
four
universities,
together,
but
university
of
baltimore
being
a
new
university,
had
to
do
a
subcontract
through
umc,
so,
which
is
why
you
just
see
three
awards
instead
of
four
okay.
A
So
the
combination
of
backgrounds
are
from
information
systems,
a
couple
of
faculties,
three
faculty
from
information
systems.
Then
there
are
a
couple
from
computer
science
at
umbc
and
then
one
from
towson
is
in
cis.
A
sema
ir
from
ub
is
actually
an
urban
planner
and
she's.
I
think
in
the
business
school,
but
she
does
a
lot
of
different
types
of
work
and
our
students
come
from
applied
math
to
applied
I.t
to
business,
schools,
backgrounds
and
sharad
is
a
computer
science
professor
at
the
bowie
state.
A
Okay,
you
can
see
a
little
bit
of
a
picture
here
and
then
there
are
different
projects
that
students
are
participating
and
have
already
produced.
Some
results
I'll
talk
a
little
bit
about
the
projects,
but
essentially
there
are
three
things
that
we
are
doing.
A
One
is
module-based
teaching
where
students
can
actually
learn
from
teaching
modules
and
I'll
talk
about
that.
The
second
one
is
through
by
getting
involved
with
our
community
in
the
baltimore
area,
in
particular.
Talk
about
that
as
well,
and
the
last
one
is
to
get
involved
in
research
projects.
So
the
three
together
we
are
hoping
for
a
complete
whole
experience
for
the
students.
A
So
this
is
a
project
structure.
We
haven't
really
put
together
the
advisory
board
yet,
but
we
are
in
the
process
of
doing
that.
Umbc
is
a
coordinating
and
implementing
organization
and
then
the
three
universities
that
I
mentioned
and
then
there
are
quite
a
few
local
organizations
in
the
city
of
baltimore,
some
private
organizations
and
and
some
government
as
well.
So
it's
a
combination
of
multiple
stakeholders-
and
there
are
in
fact
did
a
workshop
in
july
and
there
are
600
participants
in
from
the
baltimore
area
from
these
three
sectors.
A
Just
to
discuss
about
the
different
types
of
issues
that
are
can
be
solved
using
data
science
in
the
city
of
baltimore.
Our
goal
obviously
is
to
really
expand
and
make
sure
that
these
are
applicable
in
to
other
cities
across
the
country,
and
I
have
started
talking
a
little
bit
with
people
from
portland
the
portland
area,
because
they
are
also
doing
smart
city
projects.
A
But
the
hope
is
it's
going
to
expand
okay,
so
this
is
sort
of
the
types
of
things
types
of
topics
that
we
are
covering
in
the
modules
and
so
at
the
center.
We
have
the
data
science
core
and
there
are
25
students
across
four
universities
that
are
participating,
and
there
are
a
few
volunteers
who
are
not
actually
paid
but
or
cannot
be
paid
because
they
are
not
u.s.
Citizens
excuse
me,
but
they
are
also
participating.
So
around
80
is
my
guess
of
the
number
of
students.
Not
80.
A
30
is
my
guess
in
terms
of
the
number
of
students
that
are
really
participating
in
the
data
science
core,
so
the
larger
circle
right
above
the
yellow
which
is
orange.
I
believe
it
talks
a
little
bit
about
the
different
challenges
with
data
science
projects.
You
know
we
talked
to
many
different
colleagues
and
they
said
that
one
of
the
biggest
challenges
is
really
to
clean
the
data
and
get
it
into
a
format
where,
after
which,
you
can
actually
analyze
it.
A
So
the
interesting
novel,
exciting
things
are
all
in
the
analytics,
but
analytics
cannot
be
done
until
and
unless
you
have
got
the
clean
data
and
you
have
also
got
you
know
no
missing
attributes
or
more
missing
data.
If
you
have
any
you
have
to
impute
before
that,
there
is
no
point
in
trying
to
do
that
and
there
are
major
challenges
in
doing
that.
A
Machine
learning
and
deep
learning
is
a
major
part
of
this
venture
and
then
virtual
reality
and
I'll
give
you
some
examples
of
that
and
then
at
the
top
we
have
the
different
types
of
applications
so
transportation.
We
are
actually
collaborating
with
the
maryland
department
of
transportation
and
ut
state
highway.
Administration.
A
A
So
it
is
very
important
to
be
able
to
figure
out
whether
we
are
also
securing
the
cyber
infrastructure
that
we
are
building,
and
that
is
part
of
the
cyber
security
which
is
not
listed
here.
So
I
said
a
couple
of
things
on
that,
so
in
terms
of
the
education,
we
are
in
the
process
of
building
the
modules
right
now
and
the
modules
will
be
tested
on
the
cohort,
the
first
quarter
of
students
in
the
fall
of
2020
and
then
eventually
the
question
is:
how
do
we
expand
it?
A
So
we
are
planning
to
do
a
badging
system.
It's
a
special
batch
for
smart
cities,
data
science
and
smart
series,
which
doesn't
exist
right
now,
there's
a
badging
program.
We
have
nearly.
We
have
to
design
a
logo
and
all
that,
but
is
offered
by
the
university
assistant
system
of
maryland.
Okay,
so
they
are
all.
I
think
there
are
13
campuses.
A
I
think
the
vast
majority
of
the
campuses
are
participants
to
this.
Some
of
them
are
listed
here
and
the
idea
is
that
it
will
be
available
to
undergraduate
students
to
get
a
badge
essentially
so
once
they
take
the
modules
do
the
assignments
and
the
quizzes,
they
will
be
eligible
to
get
a
badge,
and
this
will
be
a
holistic
education
as
it
pertains
to
smart
cities.
A
The
other
big
piece
that
we
have
is
community
engagement,
and
this
is
not
a
video,
but
we
had
so
vinia
which
stands
for
baltimore.
Neighborhood
indicator
alliance
is
sort
of
a
center
that
works
with
all
of
the
city
councils
in
city
government,
many
of
them
at
least,
and
then
some
within
the
state
and
some
even
outside
the
state,
both
private,
as
well
as
public,
to
figure
out
what
their
needs
are.
You
know
what
are
the
major
most
major
challenges
and
then
how
can
we
actually
solve
them
with
data
science
projects?
A
And
then
how
do
we
disseminate
the
information?
How
do
we
make
the
users
actually
use
them
and
get
gather
information
from
them
about
the
usability
and
usefulness
of
those
systems?
A
So
the
user
participation
is
very,
very
strong,
a
very
strong
part
of
this
project
and,
like
I
said
there
was
a
bottom
of
data
week
in
which,
over
a
five-day
period,
we
virtual.
Obviously
this
year
there
were
as
many
as
600
participants
from
all
over
the
state
of
mainland
many
from
baltimore
city
and
some
outside
as
well,
who
came
together
and
did
a
lot
of
brainstorming.
A
We
also
had
students
part
present,
some
of
the
work
that
they
have
done
and
after
their
presentation
we
kind
of
talked
to
them,
and
then
we
had
panel
discussions
and
breakout
sessions.
We
talked
to
the
students
and
they
actually
said
that
they
learned
a
lot
of
skills
by
doing
the
projects
that
they
did
not
have
before.
A
So
this
is
the
other
thing
that
we
are
going
with
a
curriculum
map
and
on
the
one
hand,
this
is
still
a
work
in
progress
on
the
in
the
rules.
We
have
the
various
different
educational
objectives:
okay,
for
example,
critical
analysis
and
thinking.
A
You
know
written
and
oral
and
verbal
communication
technology,
competency
and
things
like
that.
So
this
is
not
a
complete
list,
and
this
is
very
much
in
a
work
in
progress
which
will
be
completed
by
this
year
and
the
columns
we
have
the
different
modules
as
well
as
activities.
So,
as
you
can
see
there
are
you
know
the
core
activities
and
projects
which
are
participating
in
the
communities
and
also
participating
in
research
projects.
A
And
then
we
are
planning
to
use
edx,
which
is
an
open
source
software
for
educational
purposes
and
it's
in
its
non-profit
and
that's,
I
think,
the
distinguishing
feature
from
coursera
for
instance.
So
we
are
planning
to
use
edx
we're
actually
going
through
the
training
right
now
to
port
our
modules
into
edx
platforms,
and
then
it
will
be
freely
available
to
certainly
to
everyone
in
the
usm
system.
But
our
goal
is
also
to
make
it
open
nationwide
and
perhaps
even
worldwide.
A
So
on
the
top
left
is
the
different
types
of
crimes,
because
unfortunately
crime
is
one
of
the
major
issues
in
baltimore
city,
but
so
a
student
actually
collected
data
and
showed
the
number
of
crimes
and
the
types
of
crimes
in
histograms,
as
you
can
see
before,
and
after
covey,
okay-
and
there
is
some-
there
are
some
interesting
patterns.
So
I
won't
go
through
the
details
right
now,
but
students
have
really
found
some
interesting
patterns
in
terms
of
the
crime
before
and
after
took
place.
A
The
one
on
the
top
right
corner
is
about
drugs
and
we're
also
extending
to
human
trafficking.
Although
human
trafficking
is
not
really
a
big
problem
in
this
region,
it
is
more
of
a
problem
in
the
water
states,
but
so
we
are
working
with
a
local.
You
know:
law
enforcement
agency
called
haida.
There
are
33
hiders
across
the
united
states,
there's
one
in
maryland
and
we're
working
with
them.
Now
these
agencies
they're
they're,
very
good
at
catching.
A
The
criminals
so
to
speak,
but
they
have
not
really
expanded
their
scope
to
social
media
and
as
social
media
becomes
more
and
more
popular
and
they're.
More
and
more
particularly
the
younger
generation
depend
depend
on
social
media
there's
a
lot
of
drug
dealing
which
is
at
the
supply
at
the
retailing
end
of
this
whole
system
that
happens
using
social
media.
So
our
plan
is
really
to
combine
the
databases
that
they
already
have
and
also
get
some
information
from.
You
know
fincen,
which
is
financial.
You
know.
A
Essentially,
these
are
money
laundering,
which
is
also
related,
because
once
drug
dealers
get
the
money
they
have
to
launder
it
to
make
sure
that
they
can
use
them
and
then
so
there
are
about
about
a
dozen
different
databases
that
will
be
connected
and
then
we're
going
to
do
some
analytics
on
that.
So
one
of
the
students
and
one
phd
student
and
several
undergraduate
students
have
already
done
some
work
in
this
particular
domain
of
trying
to
identify
the
websites
that
are
doing
drug
dealing
and
also
geolocate
them.
A
The
bottom
left
corner
shows
an
example
of
a
virtual
reality
project
that
sharma
is
doing
with
this
picture.
You
can
see
here
in
terms
of,
for
example,
virtual
reality
for
building
evacuation.
A
You
know
because
the
buildings-
and
you
know
particularly
older
cities,
have
got
designs
that
are
not
really,
you
know,
optimized
to
modern
day
living,
and
so
there
are
different
types
of
risks,
fire
and
other
types
of
things,
maybe
terrorist
at
attacks,
or
you
know,
shooting
and
stuff
like
that.
A
What
they
want
us
to
do
is
really
to
figure
out
as
early
as
possible
as
quickly
as
possible
if
a
traffic
traffic
accident
has
occurred
now
there
are
sensor
networks
with
automatic
detectors
in
various
regions,
but
they
don't
cover
everything,
and
also
some
of
the
sensors
are
unreliable.
They
go
down
malfunction
and
they're.
Not
able
to
you
know,
collect
the
data
and
send
it
to
any
local
any
remote
any
server
central
server.
A
A
So
what
we
are
trying
to
do
is
trying
to
figure
out
if
you
can
even
predict
if
an
accident
might
occur
or
an
accident
situation
is
occurring,
that
will
be
that
will
save
a
lot
of
lives
and
also,
you
know,
will
help
in
resolving
traffic
congestion
problems.
A
So
we
are
combining
multiple
modalities
data
from
social
media
data.
Coming
from
you
know,
crowdsourcing
applications
like
waze
data
coming
from
sensors,
as
well
as
from
drones,
to
see
if
you
can
put
together
a
system
that
can
combine
all
of
this
data
and
then
analyze
it,
and
then
we're
also
going
to
generate
some
data,
we're
going
to
generate
using
generative
adversarial
networks
to
create
situations
synthetically-
and
this
really
fits
very
nicely
with
the
south
data
hubs.
A
So
that's
my
oh.
The
last
thing
is
so
we
did
in
addition,
in
addition
to
obviously
the
collision,
south
data
hub
and
also
portland
state
university,
we
started
a
collaboration
recently
with
tufts
university.
This
started
in
the
nsf
pi
meeting
in
april,
where
leonard
cohen,
who
was
the
principal
investigator.
A
They
have
a
tripods
award
and
they
are
very
interested
in
working
with
us.
We
are
talking
about
student
exchange
programs
and
students
from
umc
and
towson,
and
ub
and
bowie
state
can
actually
go
over
to
staffs
some
of
the
top
students
can
come
over
to
you
and
all
of
the
other
universities
and
spend
time
in
our
labs
and
thereby
we
can
kind
of
expand,
and
we
are
learning
a
lot
of
new
challenges
that
we
really
didn't
think
about
when
it
comes
to
data
science
from
them
as
well.
A
B
Thank
you
thank
you,
and
so
I
see
that
some
people
have
asked
questions
in
the
ether
pad
and
yeah.
Thank
you
for
that
talk.
It
was
really
it's
really
well
done
and
about
the
project.
So
if
you
can
click
on
the
little
like
click
on
your
picture
and
look
at
the
little
or
click
on
my
picture
all
right
and
look
at
the
little
three
dots
that
are
next
to
it,
you
can
make
me
host
again
that
way.
I
can.
A
B
It
over
to
the
next
speaker,
but
if
you
would,
if
you
wouldn't
mind
going
in
and
looking
at
the
questions
people
have
asked,
there
were
some
notes.
Some
people
were
taking
notes
on
what
you
were
saying,
but
also
some
you
know.
Thank
you
patty.
I
transferred
your.
She
put
some
resources
in
the
chat
and
tried
to
transfer
it
into
the
ether
pad.
B
So
we'll
we'll
keep
going
and
have
the
the
next
speaker
is
dr
mantra
mantravari,
and
he
will
introduce
himself
and
also
talk
about
a
project
where
he
is
doing
a
project-based
teaching
and
what
they
are
looking
at
doing
so
I
will
transfer
the
if
you
can
unmute
and
introduce
yourself
and
then
I
will
transfer
you
the
possibility
to
share
your
spring.
C
I
am,
I
am
using
my
audio
from
my
phone.
B
C
Sure
I
am
dr
muntevadi
and
I
have
passed
a
decade
of
experience
in
online
teaching
in
data
science
and
in
big
data.
I'm
a
health
economist.
Now
we
will
be
discussing
today
about
more
of
the
real
world
applications
of
big
data
in
teaching,
as
using
as
an
example
of
the
quality
matters,
national
certification
for
online
teaching,
which
is
going
to
be
especially
relevant.
These
days
we're
trying
to
deliver
online
quality
learning,
especially
with
covet.
C
We
need
to
really
be
able
to
create
trustworthy
standards
on
expectations
of
how
exactly
to
create
the
most
effective
online
course
for,
in
this
case,
we'll
be
talking
about
healthcare
analytics
and
data
science,
especially
in
health
informatics,
and
how
exactly
we
can
provide
a
most
optimal
teaching
approach
and
different
ways
about
how
quality
matters
certification
will
help.
You
meet
that
particular
creation
of
an
effective
course
and
I'll
be
giving
some
examples
for
both
a
graduate,
as
well
as
a
undergraduate
course
in
different
ways
that
you
can
provide
more
of
an
interactive
experience.
B
C
From
a
phone
okay,
so
I
will
why
don't
I
switch
over
to
a
computer
audio
okay?
So
let
me
do
so
one
quick
second.
C
Thank
you.
Let's
see,
if
it'll,
let
me
do
that.
B
B
Okay,
so
I
think
one
problem,
maybe
that
I
think
arya
you
may
not
have
made
me
host
again
that.