►
Description
We sat down with Hyejin Yoon, Data Engineer from SOCAR - a Korean car-sharing startup - to hear about her experience with DataHub and our OSS Community.
Learn more about DataHub: https://datahubproject.io
Join us on Slack: http://slack.datahubproject.io
Follow us on Twitter: https://twitter.com/datahubproject
A
B
B
I'm
a
junior
data
engineer.
I
work
in
a
car
sharing
company
in
south
korea
called
socar,
and
my
team
is
data
platform
team.
I
am
we,
our
team
is
working
on
a
lot
of
different
projects.
For
example,
we
manage
our
own
clusters
and
we
also
develop
our
data
applications
as
well,
and
we
are
also
poc
a
lot
of
open
source
applications
as
well
and
yeah,
and
actually
I
have
been
working
here
for
six
months.
So
I'm
a
I'm
a
quite
newbie,
so
yep
great.
A
B
Actually,
the
project
was
going
on
before
I
got
on
the
company,
so
there's
a
there's,
a
strong
needs
of
you
know
introducing
a
metadata
platform
to
our
company
and
there's
this
one
senior
that
wants
to
do
it
for
so
long
like
over
one
year.
He
was
really
trying
to
do
like
this
time,
we'll
do
it
like
then
so
so
there
was
this
test
first
and
they
were
just
researching
a
lot
of
open
source
metadata
platforms
and
one
of
them
was
data
hub
and
I
think
yeah.
C
A
D
B
A
C
B
C
B
C
A
B
I
have
to
say
that
I
didn't
really
openly
publish
this
application
to
the
whole
organization.
We
just
demoed
to
few
people
and
in
the
preparation
stage.
But
we're
expecting
to
like
the
major
thing
that
we're
expecting
for
data
is
a
lot
of
people
like
a
lot
of
various
physicians
in
our
company.
Using.
C
B
To
their
daily
tasks
like,
for
example,
not
just
data
analysts
or
data
engineers,
for
example
like
pms
and
marketers,
they
all
use
that
it's
really
open.
Our
data
is
really
open,
but,
like
the
problem
is
they're
really
struggling
like.
How
can
I
find
this
data
or
like
yeah?
What
does
they?
What
does
this
column
means
like?
What
does
this
like?
Abstract
word
means
something
like
that
and
like
questions
are
really
like,
we
have.
B
We
have
our
own
slack
channel
that
answers
like
kind
of
questions
and
we're
just
answering
by
hand,
and
it's
really,
I
would
say
we-
it's
really
unproductive
because
we're
already
busy-
and
you
know
those
kind
of
questions-
can
be
repetitive
and
everything
right
and
we're
expecting
to
solve
like
at
least
nine.
Ninety
percent
of
the
questions
to
beta
health,
like
we're
just
we're
planning
to
just.
D
D
B
I
think
that
can
be
useful
because
I
don't
want
to
say
like
technical
things,
but
we're
just
solving
that
problem
into
like
using
other
platforms
called
apache
flow
and
it's
like
a
different
scheduler,
and
you
know
we're
just
connecting
like
three
other
platforms
with
it.
But
if
you
can't
just
ingest
it
on
the
ui,
that
can
be
very
helpful
and
there's
one
and
another
thing
is
a
view-based
policy
called
our
back
like
currently,
data
hub
doesn't
have
a
view-based
policy,
like
every
user
can
see
every
data
exactly.
C
A
Like
thinking
about
data
hub
as
a
whole
kind
of
like
the
current
set
of
features,
is
there
any
part
of
data
hub
that
you
think
is
like
the
most
impactful
like
either
through
use
case
or
through
future
or
through,
even
like
the
developer
experience
and
actually
like
integrating
with
it
like
what?
What.
B
And
I
picked
just
like
10
like
about
10
or
so
topics
about
wow
some
important
things,
and
you
should
consider
when
you're
deploying
metadata
platform
and
the
first
thing
was
it's
ui
like
a
good
developer,
but
like
it's
a
platform
that
everyone
in
the
company
should
use
so
like
the
ui
is
really
really
important.
I
think
I'm
a
developer,
but
like
yeah,
that's
like
when
you
when
I
like
go
to
the
meeting.
B
I
compared
to
like
the
same
operation
like,
for
example
like
in
just
like
using
the
same
data
source
and
ingesting
the
same
metadata
to
different
applications
like,
for
example,
data
and
instant,
and
the
surprising
thing
was
the
the
program.
The
script
that
is
used
by
that
is
used
for
the
metadata
condition
like,
for
example,
in
data
hub
that
that
was
just
like
10
lines
of
script.
But
when
you
do
the
same
operation
to
other
applications,
you
need
like
over
100
like
at
least
50
lines
or
like
these,
for
example.
B
D
B
I'll
say
I
love
the
ingestion
channel
most
because
I
asked
questions
there
like
the
most
and
like
the
reason
that
I
asked
the
question
on
the
channel
the
most
is
I
was.
That
was
the
most
struggling
part
for
me
because,
like
other
things,
like
the
other
part,
that
other
part
was
just
okay
like
I
could
use
helm
chart
and
I
could
use
just
some
other
things.
But
ingestion
like
on
the
ingestion
ingestion
part.
C
B
A
B
Yeah
and
and
say
first,
you
need
to
know
what,
like
what
data
sources
you
want
to
ingest
to
data
hub,
because
that
you
know,
if
you
do
that
it
just
you
know
narrow
the
scope,
a
lot
more
yeah.
There
are
plenty
of
different
data
sources
and
it's
slightly
different.
Like
the
ingest
just
methodology,
it's
a
slightly
different,
so
yeah.
If
you
choose
like,
if
you
decide
like
what
data
source
and
what
database
or
table
you
need
to
ingest
to
data
hub,
then
you'll
just
sort
of
everything.
Naturally
yeah.
C
B
I'm
thinking
more
of
a
community-wise
answer,
just
you
know,
post
a
question.
You
know
frequently
like
yeah,
someone
can
answer
it
and
you
know
feel
appreciated
for
that
and
maybe,
like
three
months
later,
six
months
later,
you
can
just
you
know,
answer
them
for
other
members
as
well,
and
that's
how
community
words
see
right.
A
C
C
B
B
B
C
A
At
home
non-stop
all
the
time
forever.
That's
when
I
started
getting
really
involved
in
more
kind
of
like
global,
or
you
know,
communities
outside
of
just
like
my
immediate
office,
right
where
it
was
just
like
it.
Just
all
of
a
sudden
felt
more
accessible.
Somehow-
and
I
don't
know
what
that
was,
but
it
was
like.
Oh
we're
all
just
at
home
together,
so
yeah
do
you
want
to
collaborate
or
what
you
want
to
like
build
something.