►
From YouTube: What is Data Science?
Description
No description was provided for this meeting.
If this is YOUR meeting, an easy way to fix this is to add a description to your video, wherever mtngs.io found it (probably YouTube).
A
Hello,
everyone
welcome
to
the
pre-session
for
the
data
science
ama.
We
thought
it
might
be
a
good
idea
that
before
we
actually
had
the
ama,
we
spent
five
ten
minutes
to
briefly
chat
with
you
all
about.
You
know
what
data
science
is
and
what
we're
doing
here
so
to
start
off
the
data
science
team,
it's
a
team
of
two.
It's
cloudy
and
myself.
A
B
Thank
you,
kevin
hi,
everyone.
My
name
is
claudia
and
I
joined
kids
lab
three
months
ago
and
last
couple
of
years
I
spent
in
a
consulting,
but
I
have
a
background
in
the
computer
science
I'm
based
in
a
melbourne,
australia,
but
you
can
figure
out
based
on
my
accent
that
I'm
not
from
australia
and
outside
work
I
enjoy
being
outside,
including
running
and
mountain
climbing
next
slide
and.
B
B
So
we
encourage
you
to
have
a
look
and
if
you
have
any
questions
just
reach
out
to
us,
it
lists
not
just
projects,
but
also
it
explains
what
kind
of
tools
we
are
using
and
how
to
set
them
up,
and
also
we
added
this
section
commonly
used
terms
in
data
science,
which
explains
some
of
the
terminology
that
we
are
using
on
a
daily
basis
and
also
we
are
encouraging
you
to
join
the
slack
channel
bt
data
science,
where
bt
stands
for
business
technology
and
yeah.
B
I'm
just
looking
forward
for
your
feedback
on
this
one.
B
So,
let's
start
with
what
is
data
science
data
science
is
a
field
that
combines
computer
science,
mathematics
and
domain
expertise
to
extract
the
insights
from
the
data.
So
we
are
using
the
best
out
of
all
of
those
words
to
not
just
understand
the
data.
What
is
there,
but
actually
sometimes
even
go
bit
further
and
ask
and
answer
some
of
the
questions
that
we
may
have.
So
some
of
the
very
common
data
science
projects
include
defining
likelihood
of
certain
events
to
happen.
B
So
some
of
the
most
common
common
questions
is
what
is
the
difference
between
data
science
and
data
analytics,
and
so
here
we
listed
some
of
the
high
level
overview.
But
what
I
would
want
to
highlight
is
the
fact
that
data
science
and
data
analytics
shouldn't
be
seen
as
a
separate
things.
They
should
be
seen
as
the
things
that
they
work
together,
so
in
data
analytics
goal
is
to
understand
the
current
state
of
the
business.
B
So
we
are
taking
the
information
from
the
past
till
today
and
we're
trying
to
understand
how
our
business
looks
like
today.
Data
science
goes
takes
this
data,
but
goes
a
step
further,
which
is
try
we're
trying
to
define
what
will
happen
in
the
future,
so
we
can
prepare
for
that.
So,
for
example,
what
is
going
to
happen
with
our
client
like?
Will
the
client
buy
our
product
or
not,
but
data
analytics
may
answer
the
question
who
bought
and
buy
how?
B
Much
so
that's
that's
one
of
the
difference,
but
again
it's
a
high
level
generalization
because
every
company
may
just
adjust
this
definition
to
their
own
needs
in
data
analytics.
B
Very
common
tools
are
data
visualization
tools
to
build
dashboards
such
as
tableau
power,
bi
and
sizes,
whilst
in
data
science
we
use
mostly
python
and
r,
which
are
the
language
computer
programming
languages
that
allows
us
to
build
models,
apply
statistical,
modeling
and
so
on.
But
if
you
have
any
questions,
just
always
please
reach
out
to
us,
and
we
are
more
than
happy
to
discuss
the
differences
and
now
passing
to
you
kevin.
A
Great
thanks
gladia,
so
one
of
the
questions
we
get
asked
a
lot
is
you
know?
How
does
data
science
differ
from
mlaps
ml
stands
for
machine
learning
you
might
have
heard
of
it.
Mlaps
is
very
buzzy
very
new
and
trendy
right
now,
and
I
think
the
best
way
to
think
about
ml
ops
is
with
this
little
sat
analogy
where
ml
ops
is
the
data
science,
but
git
lab
is
to
software
development.
A
Mlaps
is
just
a
way
to
allow
data
scientists
to
create,
deploy,
monitor,
update
our
models
and
our
predictions
in
a
more
kind
of
automated
and
streamlined
process,
so
kind
of
with
the
the
brief
landscape
of
data
science,
you
might
be
interested
in
what
what
are
we
actually
doing?
What
are
we
working
on
these
days?
The
first
major
project
that
we're
undertaking
is
something
called
parents
of
the
buy
or
propensity
to
upsell.
A
So
essentially,
what
we're
doing
is
we're
looking
at
all
of
our
existing
customers,
all
resisting
paid
customers
and
we're
using
data
science
machine
learning
to
identify
of
those
customers
which
ones
are
most
likely
to
increase
their
spend
increase
their
arr
upon
subscription
renewal
and
if
we
can
better
identify
those
customers
and
understand
the
characteristics
that
lead
to
kind
of
ar
expansion,
we
can
better
provide
the
tools
and
resources
to
the
sales
team
to
kind
of
get
get
those
customers
that
might
be
on
the
edge
kind
of
over
the
hump
to
spend
more
money
essentially
and
how
this
really
manifests
itself.
A
You
know,
so
we
build
these
predictions.
But
how
do
you
actually
use
these
predictions?
How
do
you
actually
make
sense
of
it?
There's
kind
of
it's
a
two
key
ways
where
you
can
kind
of
use
some
of
this
propensity
stuff,
so
we're
talking
about
sales
specifically
in
this
example-
and
you
know
so,
if
we
look
at
in
salesforce,
you
know,
for
example,
we
can.
A
We
can
take
these
scores,
take
these
predictions
and
we
can
import
them
into
salesforce,
so
at
the
individual
account
or
record
level
when
a
sales
rep
goes
in
to
look
at
an
account,
we
can
see
things
like
what
is
the
the
likelihood
of
this
customer
to
increase
arr,
so
likelihoods
zero
to
a
hundred
percent,
100
percent
means
they're,
definitely
going
to
do
it.
A
0
means
they're,
definitely
going
to
not
so
52
percent
might
mean
yeah,
they're,
they're,
moderate
they're
kind
of
on
the
cusp,
and
we
can
also
from
the
predictive
model,
figure
out
what
are
some
things
that
we
can
do
to
kind
of
push
that
person
along
to
kind
of
increase,
their
their
probability
of
of
of
of
increasing
their
spend
and
the
other
way.
A
This
kind
of
manifests
itself
is
kind
of
at
a
more
global
or
at
a
team
level
where
we
can
kind
of
look
at
things
in
aggregate,
so
we
can
look
at.
We
can
use
this
data
to
kind
of
forecast.
What's
what's
going
to
happen,
based
on
the
average
accounts
likelihood
to
do
something
we
can
look
at
things
like.
What's
going
to
be
their
average
predicted
upsell,
what
accounts
are
more
likely
than
others
we
can
rank
them.
A
We
can
create
different
campaigns
for
different
accounts
based
on
their
scores,
and
we
can
target
them
differently
based
on
what
we
know.
What
we
think
we're
gonna
do
and
that's
kind
of
the
end
goal
is
to
take
all
this
information
and
to
score
every
single
paying
account
with
this
model.
So
we
can
predict
what
we
think
they're
going
to
do
in
the
future,
so
some
pretty
cool
stuff.
We
have
a
lot
of
other
stuff
planned.
A
First
and
foremost,
we
have
a
lot
more
predictions
that
we're
looking
to
do
so
that
was
printed
to
buy
or
upsell
we're
also
looking
at
churn.
So
can
we
predict
who's
likely
to
churn
totally
as
a
customer?
You
know
to
not
subscribe
anymore,
maybe
to
reduce
their
subscription
lead
scoring
so
converting
free
to
paid
customers,
maybe
looking
at
how
we
convert
people
from
a
trial
to
a
paid
customer
and
then
looking
at
kind
of
the
the
actual
dollar
amount
for
expensive
buy.
A
So
can
we
predict
yes,
this
person's
going
to
increase
their
error,
but
can
we
actually
tie
a
dollar
onto
that?
Let's
say
yeah
they're
going
to
increase
it
by
two
thousand
dollars.
A
Another
big
area
is
customer
segmentation,
so
using
machine
learning
to
kind
of
find
these
natural
groupings
of
customers
based
on
kind
of
various
characters,
characteristics
about
them
and
see
how
they
see
how
they
naturally
group
together-
and
we
also
have
the
the
ml
offset.
I
mentioned
before,
working
with
with
product
to
help
kind
of
define
what
what
ml
ops
is,
what
features
we
would
like,
as
some
in
that
product
we
could
actually
use
that
product.
A
We
can
kind
of
dog
food
and
use
our
own
models
to
to
beta
test
things
out
and
and
hopefully
develop
a
product
that
that
we
really
like
using
as
well
and
finally,
one
thing
that
I'm
really
interested
in
is
building
out
what
we're
calling
a
data
science
tool
set
so
cloud
by
myself.
We
don't
want
to
be
the
only
two
people
in
the
entire
company
who
can
build
predictive
models.
A
We
want
to
create
tools
and
enablement
to
allow
anyone
in
the
company
who
wants
to
build
predictive
files
to
build
protective
models,
and
so
we've
developed
kind
of
a
framework
and
a
process
and
some
tools
to
allow
you
to
do
that.
If
you
want
to
do
that
so
kind
of
the
first
step
of
that
is
we
created
what
we're
calling
a
data
science
environment?
A
The
link
is
in
the
slides,
and
this
is
kind
of
a
it's:
a
jupiter
environment,
jupiter,
notebook
environment
that
has
all
the
common
python
modules
and
tools,
kind
of
set
up
and
ready
to
go
and
connect
it
to
our
snowflake
environment,
where
all
the
data
is
housed.
So
it
really
allows
you
to
kind
of
get
up
and
going
quick
and
to
to
start
doing
some
analysis
or
data
science
without
having
to
spend
a
lot
of
time
trying
to
figure
out
how
to
set
everything
up,
and
we
always
welcome
contributions.
A
If
you
have
an
interesting
idea
or
a
module
or
a
package
that
you
would
like
to
see
in
it
feel
free
to
contribute,
feel
free
to
drop
us
a
note.
I
think
we're
always
pretty
receptive
to
seeing
how
we
can
improve
some
of
that
stuff.
B
Thank
you
kevin.
Thank
you
so
much
for
introducing
us
to
concepts
of
a
project
to
breaking
those
to
our
audience
and
your
audience.
Thank
you
so
much
for
watching
and
if
you
have
any
questions,
just
please
reach
out
to
kevin
or
to
myself
and
we're
more
than
happy
to
answer
any
data
science
and
non-data
science
questions
you
might
have.
Thank
you
so
much
thanks.
Bye.