►
Description
In this public meeting, Joanne Cheng presented some ideas and practices for drawing geographical maps using Clojure and R.
It was a special meeting since we explicitly invited R users who are new to Clojure to join us and listen. We plan to have more meetings of this kind in the future.
Moderator: Jake Nylund
Text chat:
https://tinyurl.com/y789wftk
A
Well,
I
guess
without
any
further
ado
joanne,
would
you
like
to
take
it
away.
B
B
I
guess,
whichever
is
comfortable
for
you
yeah.
I
guess
I'll
just
get
started
yeah.
If
anyone
has
any
questions
or
ever
wants
to
stop
me,
while
I'm
talking,
please
feel
free
to
do
so
yeah.
So
today,
I'm
just
going
to
be
talking
about
sort
of
my
exploration
with
making
creative
maps
with
closure
and
r.
B
So
a
little
bit
about
me,
my
name
is
joanne.
I
am
the.
I
am
a
front-end
developer
at
pitch,
which
is
also
what's
powering
this
presentation.
Right
now.
I
work
in
team
data
visualization.
We
don't
have
maps,
but
we
are
in
charge
of
helping
people
make
charts
and
graphs
for
their
presentation
yeah.
B
I
have
some
links
to
my
twitter
and
my
github
if
you're
ever
interested
or
need
to
ping
any
questions
that
you
don't
have
a
chance
to
ask,
so
I
guess
a
little
bit
about
me
and
my
hobbies.
I
I'm.
A
I
think
we
were
getting
a
little
bit
of
feedback
there,
so
I
I
did
a
terrible
job
of
this,
so
I
can
interrupt
you
real
quick.
If
we
can
ask
if
you
guys
do,
have
a
question
put
it
into
chat
first
and
then
I'll
try
and
interrupt
at
a
convenient
point
for
you
as
you're
speaking
joanne
cool.
B
No
race,
so
I
I've
been
working
in
database
for
a
couple
of
years,
mostly
in
front
of
development,
and
some
of
my
hobbies
include
making
personal
maps
for
hikes
and
trips
that
I
I've
done
over
the
years
and
I
really
enjoy
maps
because
I
think
they're
a
really
great
way
to
tell
a
lot
of
its
stories
and
information
in
just
one
picture.
So,
for
instance,
here
you
can
see
a
correlation
between
heat
sort
of
the
temperature
of
a
neighborhood
and
income.
B
So
this
was
part
of
an
article
about
how
poor
neighborhoods
in
the
u.s
tend
to
be
more
affected
by
climate
change.
B
So
I
I
come
from
this.
I
come
at
this
and
I'm
talking
about
this
through
more
of
a
hobbyist
perspective,
I
really
enjoy
making
maps,
but
I
am
not
a
just
pro
I'm,
not
a
gis
professional,
and
I
also
really
wanted
a
break
from
web
development
in
my
spare
time
since
I
deal
with
all
the
fun
stuff,
all
the
fun
issues
of
the
browser.
So
that
is
why
I'm
giving
this
presentation
today
and
sharing
why
I
chose
r
instead
of
using
web
mapping.
B
So
this
is
also
kind
of
a
shortened
version
of
the
talk
that
I
gave
in
closure
d,
which
is
posted
on
the
internet
somewhere
on
youtube,
where
I
basically
kind
of
dove
into
this
world
of
making
maps
with
our
enclosure
because
of
my
homesickness
from
moving
to
berlin
last
year,
but
something
I
noticed
while
sort
of
making
some
of
these
maps
with
our
and
closure
is
that
there's
just
a
lot
of
resources
out
there.
B
So
one
of
my
goals
of
this
talk
is
is
sort
of
just
distill
all
those
resources
into
one
place.
So
all
of
you
have
a
good
starting
point
to
create
your
own
maps
and
maybe
a
good
reference
for
yeah.
When
you
start
diving
into
this
yourself,.
B
B
So
I
decided
on
r,
because
the
language
has
a
lot
of
libraries
to
decode
and
read
geographic
file
formats
and
there's
a
lot
of
libraries
for
drawing
data
like
ggplot2,
and
I
kind
of
like
the
way
that
it
kind
of
works
well
with,
like
my
brain
after
writing,
closure
for
the
last
couple
years.
B
It's
I
really
like
how
everything
is
treated
as
a
list
and
since
you're
working
with
collections
of
data,
I
think
that's
a
really
good
way
to
kind
of
think
about
programming,
so
it
it
wasn't
too
much
of
a
it.
Was
nice
to
be
able
to
transfer
some
of
my
knowledge,
so
r
can
also
make
really
beautiful
maps
in
charts.
B
B
I
think
I
mixed
up
some
of
my
slides,
so
I'm
gonna
improvise
a
little
bit,
so
this
was
actually
from
will
cohen's
talk.
He
gave
the
first
talk
on
this
sort
of
gis
and
in
this
sort
of
closure
data
science
series.
B
So
some
of
the
trickiest
parts
I
found
when
working
with
js
data
was
just
kind
of
dealing
with
all
these
formats,
so
in
will's
talk,
he
talks
a
lot
about
how
to
deal
with
things
like
how
like,
for
instance,
like
post
gis
lets
you
perform
all
these
spatial
operations
on
these
shape
files
and
sorry.
This
geospatial
data
and
r
gives
you
the
same
capabilities
to
interact
with
the
same
the
same
software,
but
the
frustrating
parts
and
why
I
started
diving
into
closure.
B
B
But
there's
a
lot
of
strange
rules
that
I
kind
of
didn't
really
grasp
and
understand,
even
when
going
through
examples
and
tutorials,
and
I
also
found
it
difficult
and
maybe
a
little
annoying
to
do
seemingly
easy
things,
especially
coming
from
closure,
which
is
why
I
started
diving
into
clue.
Juicer.
B
So
clogister
is
a
closure
library
that
talks
to
r
using
an
r
library
called
rserv
or
also
there's
other
methods
that
clogister
supports
right
now,
but
I
tend
to
stick
with
rserve,
so
some
really
cool
things
about
clojusure.
B
It
lets
you
call
our
libraries
within
your
closure
code
and
treat
them
as
and
you
can
kind
of
work
with
them
as
basically
like
you're
working
with
closure
external
closure
namespaces.
It
also
lets
you
use
key
value
hash
map
in
the
syntax,
which
makes
things
a
little
bit
cleaner.
As
you
can
see
in
the
r
section,
there's
yeah
there's
just
a
lot
more
extra
text,
which
kind
of
gets
it
makes
it
easier
to
make
mistakes,
and
so
also
it
supports
sort
of
the
ggplot
syntax.
B
So
something
that
the
plotting
library
in
r
does
it.
It
overrides
the
plus
symbol
and
lets
you
basically
build
visualizations
line
by
line
like
you're,
adding
layers.
So
here
you
have
a
ggplot
object,
you're
initiating
a
plot,
and
you
can
add
layers
to
it
using
the
plus
symbol
and
so
yeah.
The
clojuster
is
able
to
support
that.
B
So
it
really
does
feel
like
you're
writing
r.
With
a
little
bit
with
a
few
small
differences.
B
So,
switching
gears-
I
guess
I
really.
I
would
like
to
show
you
how
I
work
with
this,
but
I
guess
before
diving
into
this
I'll
be
focusing
a
little
bit
on
shapefiles.
B
So
so
shapefiles
are
polygons
points
and
lines
and
also
vector
files
essentially
or
sorry.
I'm
I'm
gonna
just
take
a
little
bit
of
a
break.
B
So
there
are
vector
values
so
think
of
shapes
lines
and
points.
And
yes,
the
last
talk
given
at
cyclos
does
a
much
better
job
of
explaining
this.
B
So
the
shape
file
also
has
a
projection
which
tells
whatever
library
program
how
to
translate
these
numbers
from
coordinates
from
numbers
in
into
2d
values
and
how
they
map
onto
a
globe.
B
B
I
require
a
few
a
few
libraries
and.
B
A
I
have
a
shape
file
that
I
downloaded
from
an
open
data
site
in
new
jersey,
which
is
where
actually,
I
am
right
now.
B
B
So
I'm
requiring
a
couple
libraries.
B
B
So
sorry
about
that,
so
I'm
just
reading
in
this
file
in
these
three
lines
and
then
passing
it
to
a
function
called
st
read,
which
then
loads
this
to
what
is
called
a
data
frame
in
r
so
here
base
slash
class
is
a
base,
is
basically
ours
like
global
scope
or
basic
functions
and
class.
Just
tells
you
the
class
name
of
your
whatever
argument
is
there,
so
I
also
use
the
summary
function
to.
B
So
then
you
can
start
once
you
get
your
data
into
r
load
it
in
like
this.
You
can
start
asking
questions
so,
for
instance,
what
are
the
columns
that
are
out
there?
So
I'm
calling
our
so
this
is
still
dealing
within
the
data
frame.
B
Sometimes
when
I'm
working
with
this,
it's
sometimes
easier
just
to
deal
with
closure
types.
So
using
a
clogister
lets
you
transform
r
into
or
translate
r
into
closure
lists
and
closure
types.
B
A
A
B
So
I
guess
I'm
trying
to.
I
don't
know
why
it's
just
blinking
on
me,
but
it's
basically,
I
think
the
standard
functions
that
are
available
in
r,
so,
for
instance,
so
this
I
have
our
open
right
now.
So
this
would
these
are
the
same
function
and
I
can
do
or
use
class
as
well.
B
A
A
B
A
B
Just
kind
of
inspecting
the
data
right
now,
but
I
guess
yeah
I
should
actually
just
plot.
Oh
my
gosh.
I
completely
skipped
a
step
in
my
presentation.
I'm
really
sorry!
I
I
haven't
given
a
talk
over
a
zoom
before
and
it's
I'm
going
to
be
completely
honest.
B
It's
very
disorienting
because
there's
just
like
not
a
lot
of
like
human
interaction
I
can
see,
which
is
why
I'm
just
a
little
strange
right
now
so
usually,
actually
what
I
do
when
I
get
a
shape
when
I
download
a
shapefile,
is
I
open
in
a
program
called
qgis.
B
B
And
I
am,
I
think,
I've
just
forgotten
how
to
use
a
computer,
so
what
I
can
do
here
is
to
sort
of
get
like
a
visual
overview
of
what
I,
the
data
that
I
downloaded
and
qgis
is
a
really
really
great
open
source
program.
That
is
like
a
a
gui
interface
for
a
lot
of
these
gis
software,
which
I
I
have
couple
slides
that
will
explain.
B
I
was
trying
to
do
this
thing
where
I'd
show
the
example
first
and
then
explain
it,
but
I
don't
think
it's
now
that
I
think
about
it.
I
don't
think
it's
working
as
well,
so
what
I
do
is
usually
I'll
like
take
a
look
at
the
shapefile
that
I
have
just
kind
of
get
a
good
overview
of
it,
and
then
I
can
go
into
art
and
then
start
kind
of
playing
around
with
it.
B
So
I
had
an
exercise
where
I
wanted
to
see
this
returns
all
of
the
water
in
the
state
of
new
jersey.
But
what
happens
if
I
just
want
to
plot
the
lakes
and
just
want
to
look
at
the
lakes,
so
what
I
usually
need
to
then
do
is
find
what
column
would
I
find
the
lakes
or
what?
What
kind
of
water
or
what
column
would
I
look
for?
So
if
I
go
back
to
this.
B
B
B
Is
this?
Is
this
better?
B
A
B
I
I
I've.
Actually
I
thought
I
was
just
going
to
run
examples
and
I
didn't
have
to
do
any
typing,
but
now
something
weird
is
happening.
B
A
B
A
That's
getting
set
up,
maybe
you
can
kind
of
talk
me
through
the
data
exploration
bit
of
this.
So
sure
one
of
one
of
the
things
that
I
got
lost
on
here
is,
as
as
you
were
looking
through
some
of
these
base
functions
and
I
think
they
were
it
looked
to
me
like
you
were
trying
to
pull
out
what
the
the
specific
data
was
for
water.
A
I
forget
where
that
was
in
the
file.
Is
your
workflow
normally
that
you
will
explore
the
data
first
or
visualize
it
first,
because
to
me
I
saw
you
pulling
the
data
up
and
I
was
looking
through
and
I
I
think
I
was
talking
about
this
a
little
bit
earlier
with
some
of
the
people
on
this
call.
But
I
I
don't
read
data
like
that
cleanly.
I
have
to
see
it
first
to
go.
Oh,
this
is
what
I'm
looking
at
so
sorry.
I
can.
B
B
No
actually,
I
I
do
need
to
visualize
it
a
little
bit
first,
so
actually
I
can
just
probably
run
the
code
to
so.
If
I
were
to
actually
plot.
B
B
Yeah,
so
I
sometimes
do
also
yeah.
I
try
to
do
the
visualization.
I
have
to
get
something
out
there.
So
usually
that's
actually
why
I
use
qgis,
because
I
can
just
do
it
in
two
clicks
and
then
I
try
to
use
closure
for
the.
I
guess
other
kind
of
exploration,
but
I
yeah.
B
I
guess
it
varies
depending
on
how
much
I
need
to
visualize,
because
sometimes
I
dive
into
the
visualization
first
and
then
I
realize
I've
gotten
like
I've
gone
too
far
and
I
like
get
lost
and
trying
to
create
this
good
graph
when
I'm
not
really
sure
what
I'm
doing
so
yeah
it's
it's
kind
of
it
varies
but
yeah.
My
example
is
yeah.
Everything
is
loaded
and
everything's
working
again.
A
B
A
So
we
had
a
question
about
some
of
your
development
flows.
How
do
you
find
your
your
interactions
here
between
emacs
and
you
said
you
kind
of
go
back
and
forth
between
our
studio
and
qgis,
and
then
this
is
our
graphics.
That
seems
like
kind
of
a
lot
of
tools
to
get
what
looks
like.
B
B
B
I
think
there's
a
there's
a
couple
examples
down
below
where
I
actually
just
call
our
code
straight
from
closure,
because
there's
a
couple
of
a
couple
of
things
that
I
haven't
really
figured
out
how
to
do
in
closure.
But
it's
the
clogister
library
is
still
in
development
and
there's
a
lot
of
stuff
that
I
feel
like
sometimes
like.
B
If
I
I
can
do
fast
food
because,
like
a
lot
of
the
the
stuff
is
just
reading
files
and
yeah
formatting
strings
and
things
like
that,
in
which
case
closure
would
be
faster.
A
B
Sure
yeah,
so
I
guess
I
should
I
had
a
couple
slides
explaining
what
I
just
did
so
maybe
that
would
help.
B
Is
or
how
these
things
are
plotted
is
through
a
couple
r
libraries,
so
the
sf,
the
simple
features
library
is
responsible
for
helping
me
read
the
file
and
then
turn
it
into
a
data
frame
and
ggplot2
is
the
library,
the
plotting
library
that
you
see,
there's
also
a
bunch
of
other
really
important
data
science.
B
Libraries
that
you
download
in
this
one
package
called
tidyverse,
which
is
a
collection
of
lots
of
data
science,
libraries
which
includes
ggplot2,
something
that
s
the
simple
features
library
does,
as
it
also
adds
a
couple
of
functions
to
the
gg
plots
to.
B
Certain
name
space
so
we're
I'm
so
sorry
this
was
a
better
idea
in
my
head.
Then
I
really
wanted
to
try
something
new
with
like
live
coding
examples
and
it
turned
out
a
little
bit
messier
than
I
thought
it
would
be
so
there's
a
couple
of
functions
that
it
adds
to
the
ggplot2
namespace
to
help
you
kind
of
create
these
plots
layer
by
layer
using
this
sort
of
vector
data.
B
So
the
software
that
is
kind
of
powering
the
simple
features
and
gis
sort
of
yeah.
The
the
gis
tools
is
they're
listed
here.
B
So
gdell
helps
you
read
and
write
these
like
vector
and
raster
geospatial
formats,
such
as
the
shape
file,
the
geos
library,
helps
deal
with
spatial
types.
B
I
guess
I
also
want
to
explain
a
little
bit
about
the
data
frame
since
some
people
here
don't
have
a
lot
of
our
knowledge,
so
it's
a
2d
data
structure,
rows
and
columns,
and
it's
used
pretty
frequently
with
our
for
data
science,
it's
kind
of
like
their
key
data
structure,
and
so
the
data
we're
working
with
that
data
that
we're
importing
from
the
shapefile
gets
converted
into
a
data
frame.
So
we
can
apply
all
these
sort
of
standard
r
functions
onto
it.
B
B
So
I
guess
an
example
I
wanted
to
run
through
is
filtering
out
just
the
lakes
and
just
displaying
the
lakes
on
this
plot.
B
So
this
is
the
column
that
we
want
in
our
new
jersey,
water
data
frame
to
to
figure
out
which
what
kind
of
water
body
is
it.
So
if
we
run.
B
So
here
the
first
line,
I'm
using
a
clojus
or
functions
the
dollar
sign
to
help
me
just
get
all
the
unique
types
of
water
in
the
data
frame
I
like
to
sort
of
think
of.
B
I
should
have
I
sort
of
so
in
r.
This
would
look
something
like.
B
I'm
sorry
my
mind's
flanking,
but
I
like
to
think
of
this
as
sort
of
the
get
syntax
in
closure.
B
And
so
I
wrote
a
helper
function
to
help
me
filter
by
whatever
type
that
we
pass
it
in.
So
if
I
were
to
just
actually
I'll
just
run
the
example.
B
B
So
we've
plotted
everything,
that's
a
lake
and
then
you
can
apply.
You
can
continue
to
run
functions
on
the
plot
to
kind
of
add
more
layers.
So
this
chord
sf
is
actually
a.
B
So
this
applies
a
projection
of
the
first
shapefile
to
the
plot
to
to
the
pl
sorry,
it
basically
takes
the
projection
of
the
shape
file
and
applies
it
to
the
plot,
and
if
you
want
to
get
rid
of
the
grid
lines,
the
theme
void
function
is
really
helpful.
B
B
Yes
and
there's
a
lot
of
lakes
yeah
and
it's
I
I'm
like
now,
trying
not
to
tweak
tweak
it
and
just
I
need
to
keep
going
with
my
presentation.
B
B
So
I
am
working
on
an
example
to
get
this
working
in
clogister,
so
I
just
wanted
to
explain
sort
of
this
concept
so
there's
this
concept
in
data
visualization
called
small
multiples
where
you're
plotting
over
the
same
axes:
you're
plotting
related
data
over
the
same
axes
and
it's
rather
than
cramming
everything
into
one
plot.
B
You
break
things
up
into
this
grid
where
you
can
easily
compare
different
like
the
same
visualization
over
different
parameters,
and
essentially
I
wanted
to
apply
that
to
or
use
that
technique
for
the
sort
of
data
set.
These
links
in
new
jersey.
B
A
So
forgive
the
interruption
here:
what
from
from
the
visualization
perspective
of
this,
because
that
small
multiples,
as
you
were
talking
about
it,
that
I
can
see
absolutely
why
I
would
want
potentially
to
visualize
that,
because
I
can
see,
I
think,
from
what
I
was
seeing,
the
gray
in
the
background
and
then
the
the
blue
there
right.
A
B
Not
a
map
like
it's,
I
guess
different.
I
mean
I
like
to
compare
sizes
of
the
lakes,
so
you
you
know
looking
back
at
this.
I
guess
you
can
see
what
is
the
biggest
lake,
but
it's
kind
of
hard
to
get
a
good
picture
of
what
they
all.
I
I
don't
know.
I
think
it's
pretty
cool
to
kind
of
see
like
all
these
likes
like
in
a
different
view.
I
guess
I'll
just
run
the
example
and.
C
This
is
a
really
time
consuming
thing
to
do
it
in
qgis,
mainly,
you
would
have
to
make
a
bajillion
little
maps
all
the
same
scale
and
export
them
out,
but
if
you're,
comparing
in
this
case
lake
size
or
city
size
or
building
size
or
just
things
from
shape
to
one
another,
people
always
like
being
able
to
see
different
locations
that
are
in
disparate
places
at
the
same
scale,
is
like
a
very
powerful
visual
tool
to
realize
what
one
thing
is
for
the
other.
So
I
think
this
is
super
neat.
B
This
is
not
working.
A
A
Maybe
so,
while
you're
doing
that,
I
can
direct
this
at
will
so
that
it
sounds
to
me
like
this.
Small
multiples
is
vaguely
I'm
much
more
familiar
with
textual
data,
but
it
sounds
vaguely
like
kind
of
a
heat
map
or,
like
a
word
cloud
type
of
visualization,
where
you
can
kind
of
see
everything
in
in
a
bag
of
data.
Almost
is
that.
C
Kind
of
there's
a
reference
that
you
should
look
up,
there's
two
of
them
well,
there's
two
sort
of
there's
a
bunch
of
books
by
a
guy
named
edward
tufte,
one's
called
the
visual
display
of
quantitative
information
and
the
other
one
is
called
envisioning
information,
and
in
that
you
know
one
of
the
principles
he
talks
about
for
good
graphic
excellence.
Is
this
idea
of
small
multiples
that
if
you
hold
everything
constant,
except
for
one
thing
in
this
case,
you
know
joanne's
telling
us
about
like
the
shape
of
lakes
right
or
in
the
example?
C
She
was
just
showing
the
exact
wiggle
of
that
line.
Then,
once
you
understand
one
of
the
multiples,
all
the
other
ones
come
for
free,
so
it
saves
the
viewer
of
having
to
like
reinterpret
each
one
of
these
links
every
single
time.
So
you
know
doing
it,
for
graphs
is
one
thing
which
r
does
but
she's
like
showing
us
this
way
of
applying
small
multiples,
not
only
to
graphs
in
the
way
that
r,
like
very
classically,
does
with
ggplot
or
whatever
else,
but
doing
it
for
maps
too.
C
B
A
So
do
you
are
you
sure
that
you
have
a
screenshot
of
it?
Because
if
you
could
like
take
the
time,
if
nothing
else,
to
show
us
that
that
would
be
pretty
cool.
B
I
have
the
rf
version
of
it
actually
working
for
a
different.
This
is
for
colorado
for
a
different
state,
but
yeah.
I
guess
my
sort
of
excitement
about
this
is
being
able
to
kind
of
compare
all
these
different
lakes.
In
one
view
like
this,
instead
of
yeah
like
having
your
eyes
like
trying
to
search
around
like
a
map
for
all
these
different
things,
you
can
make
a
big
comparison
in
just
kind
of
one
s.
B
One
look
and
I
I
also
think
yeah
it's
just
kind
of
like
the
based
on
like
the
original
source.
B
I
think
it's
just
a
really
pretty
way
of
looking
at
yeah
looking
at
geographic
data
and
yeah,
like
you,
never
really
get
to
see
sort
of
all
these
things
next
to
each
other,
like,
for
instance,
if
you're
yeah
like
taking
all
these
lakes
like
in
different
corners
and
just
like,
taking
away
their
location
and
just
putting
them
next
to
each
other,
it's
kind
of
like
you're
yeah
like
just
rearranging
stuff
but
they're,
real
life,
things
and
they're
very
big,
and
you
can
never
do
this.
B
I
don't
know,
I
think
it's
pretty
cool,
but
I
can
kind
of
explain
the
process
that
I
made
to
or
process
that
I
did
to
make
this
in
the
closure
code.
B
Cool
so
unfortunately
it's
not
running,
as
I
had
hoped,
but
the
way
so
you
start
off
with
your
data
frame,
so
my
list
of
new
jersey
bodies
of
water,
you
filter
out.
B
B
So
now
you
have
just
a
list
of
all
the
lakes
and
ponds
in
new
jersey,
but
now
they're
kind
of
in
all
this
different
order.
I
think
the
most
interesting
parts
of
it
are.
The
interesting
parts
are
the
biggest
ones.
So
let's
say
you
just
wanted
to
look
at
the
20
biggest
lakes,
so
I
actually
couldn't
get
this
working
with
clojuster,
so
danielle.
B
I
may
have
to
talk
to
you
about
this
later,
but
I
ended
up,
but
even
though
I
couldn't
get
it
working,
this
was
one
line
in
r,
so
I
created
a
function
in
inclogester
that
I
created
an
r
function
in
glochester
that
I
can
call
from
closure.
So
here
I'm
ordering
by
this
column
called
shape
area,
which
is
just
the
yeah
the
area
of
the
polygon.
B
I
set
the
number
of
lakes
that
I
wanted
to
see.
Yeah
I
filtered
them
out.
So
I
applied
the
ordering
function.
B
So
I
used
a
couple
of
clogister's
helper
functions,
so
this
would
probably
look
like
a
data
frame
bracket
1
2
20.,
so
this
line
here
you're
just
getting
the
first
20
rows
of
your
ordered
list
so
yeah.
This
is
what
it
would
look
like
in
our
there's,
probably
cleaner
ways
to
do
this
with
threading.
I
think
this
worked
all
right
for
for
now
so
yeah
this
one
was
able
to
run
so
now.
B
B
B
Closure
just
imax
just
sort
of
exploded.
B
A
So
what
while
you're
getting
in
there,
I
I
want
to
apologize
to
your
beautiful
vintage
map,
because
I
think
I
was
thinking
about
it.
The
wrong
way.
A
I
I
I
was
imagining
it
more
as
like
a
navigational
you
say
map
I
think,
navigate
like
this
is
how
I
get
from
here
to
here,
and
I
I
learned
a
lot
already
in
between
you
and
will
that
how
I'm
thinking
about
map
data
was
wrong.
So
thank
you
for
teaching
me
something
the
what
you're
you're
back
up.
So
I
apologize
for
the
interruption,
go
ahead
and
keep.
B
Worries
no
worries,
but
I
used
I
I
guess
the
biggest
thing
here
is.
I
used
a
library,
an
art
library
called
calplot
to
help
me
plot
this
out,
and
this
was
still
a
work
in
progress
in
translating
this
to
closure
because
it,
but
basically
I'm
taking
every
row
of
yeah.
B
Oh
gosh,
I'm
a
little
bit
flustered
still,
but
I
created
a
plot
with
every
the
top
20
biggest
lakes,
and
then
I
found
their
centers
using
the
sf
library's
centroid
function
and.
B
Now
it's
connecting
earth.
I
just
started
up
the
session,
so
it
froze
a
little
bit,
but
I
can
keep
talking
so
with
the
center
that
becomes
like
the
center
of
the
new
plot
that
we're
creating.
So
that
way
it
can
kind
of
figure
out
what
the
boundaries
are.
B
B
B
I
also
have
this.
All
of
this
is
all
on
github.
If
you
want
to
copy
and
try
to
run
these
yourself,
I
think
I
think
I
wrote
much
better
comments
than
I
am
and
I
think
I'm
better
at
writing
things
than
I
am
explaining,
but
I'm
happy
to
answer
any
questions.
B
Which?
Which
question.
A
Chat
about
edward
tufte,
and
then
I
would
guess
the
small
multiples
would
be
ordered
similarly
by
polygon
size.
B
I
think
I'm
missing
some
context.
I.
B
Yes,
there's
will
cohen's,
yes,
that
was
actually
another
inspiration.
I
think
there
was
also
another
if
you
click
on
the
okay,
I'm
showing
it
on
my
I'm
sharing
my
screen.
So
this
is
a
hand-drawn
map
where
all
the
rivers
are
straightened
out
and
yeah.
It
is
it's
hand-drawn
and
it's
also
exaggerated.
B
This
is
also
another
thing
where
people
sort
of
take
artistic
artistic
liberties
with
sort
of
geographic
representations,
so
yeah,
also
when
like
something
else
to
keep
in
that
I've
sort
of
have
with
like
stuff
that
I
do
like
my
personal
hand-drawn
maps
like
all
of
this
is
all
up
for
interpretation.
B
So
actually
I
have
another
example
that
I
didn't
want
to
get
into,
because
I
don't
think
it
was
really
relevant
about
simplifying
lines.
So,
for
instance,
if
you.
A
Yes,
but
before
you
go
on
too
much
further,
I'm
intrigued
by
your
your
comment
about
open
to
interpretation,
and
so
my
understanding,
I
do
have
a
little
bit
of
experience
with
gis,
but
there's
certain
ways
of
representing
the
data,
but
I
wouldn't
necessarily
say
it's
open
to
interpretation
right
like
the
the
the
lake
or
the
rivers.
I
guess
in
that
case
are
what
they
are
and
now
you
can
use
different
projections
to
represent
them
visually,
but
can
can
you
elaborate
a
little
bit
on
how
you
mean
open
to
interpretation.
B
I
guess
I
want
to
when
you're
asking
that
question
some
an
example
pops
to
my
head
about
these
election
maps.
This
is
yeah.
I
think
this
is
a
pretty
good
example,
where
you
have
things
that
are
kind
of
geographically
in
the
right
place
and
you
can
see
the
shape
and
it
looks
like
the
us,
but
this
is
something
where
the
shape
is
open
for
interpretation,
and
for
this,
like
this,
this
sort
of
example
lets
you
see
like.
B
Yeah
lets
you
see
the
data
more
clearly,
rather
than
seeing
like
a
more
literal
interpretation
of
a
map,
whereas
here
yeah.
This
is,
if
you
drew
everything
point
by
point
and
geographically
correct.
You
kind
of
it's
harder
to
compare
the
length,
so
things
aren't
maybe
stretched
they're
not
stretched
out,
but
they're
just
kind
of
straightened
a
little
bit.
So
you
can
kind
of
compare
it.
You
can
compare
the
lengths
of
each
river.
B
Does
that
sort
of
answer.
Your
question.
A
The
distances
have
been
calculated
correctly
right,
so
you
know,
based
on
whatever
projection
on
a
map,
a
river
can
appear
much
longer
than
it
actually
is,
or
it
can
appear
to
be
its
actual
size.
So
all
of
that
has
been,
I
don't
know,
has
been
calculated
and
sorted
out
for
this,
and
then
it's
the
open
to
interpretation.
Part
of
this
is
it's
kind
of
been
straightened
out
into
a
a
line
to
represent
how
it
would
how
its
length
would
be
compared
to
the
the
river
next
to
it.
B
Cool
were
there.
B
Oh
jesus
gomez
wrote
a
question
about
changing.
The
l
apply
to
something
closure
native,
like
map
and
no
you
you
can't
use.
Unfortunately,
you
can't
use
closure
native
functions
on
data
frames.
A
Frame
is
that
a
normal
interaction
if
you're
going
between
you
know
more
than
our
workflow
to
using
our
enclosure
that
you
just
kind
of
mash
it
out
to
say,
hey
put
this
data
frame
into
closure
native
stuff?
Do
my
operations
mash
it
back
into
the
our
data
frame.
B
I
I
would
say
it's
probably
not
performance,
and
but
I
I
don't
know,
if
there's
really
a
normal
workflow
for
this,
because
it's
still
kind
of
in
development
and
still
a
bit
experimental.
A
I
was
curious
more
as
to
your
workflow
what
you
do
with
it
is
it.
You
know
you
just
say
hey,
I.
I
can't
work
with
your
r
stuff
here.
Give
it
to
me
enclosure
all
right
cool.
Do
my
stuff
get
it
processed.
How
I
want
all
right
here
you
go
arc.
This
is
now
what
I
want,
but
I
think
daniel
may
be
the
the
person
to
ask
a
more
intelligent
question
about
that
than
me.
B
Yes,
that's
it's
like.
I
think
I
I'm
just
a,
I
guess,
a
user
of
this
library.
I
think
daniel
has
more
insight
or
a
better
person
to
ask
about
this.
A
B
B
I
here
we
go.
B
This
is
strange.
Okay,
I
guess
my
examples
are
not
working
right
now
I
might
have
overwritten
them,
but
I
I
did
sort
of
have
an
example
about
simplifying
lines.
So,
for
instance,
you
may
not
want
to
have
this
sort
of
fine
detail.
B
Sometimes
you
just
want
like
a
visual
representation,
and
I
guess
it
kind
of
goes
back
to
that
election
map.
I
showed
before
where
you
have
just
sort
of
an
abstract
shape
and
you
have
a
general
location
of
where
it
is,
and
that's
that's
good
enough,
and
it
sometimes
is
a
good
thing
because
you
can
actually
like
when
things
are
yeah
like,
for
instance,
with
election
maps
like
when
you
see
like
a
big
blob
of
red
or
blue.
B
You
your
mind
thinks
okay,
this
is
bigger,
so
thus
this
has
more
value,
and
so,
when
you
sort
of
take
away
some
of
the
detail,
you
can
describe
other
stuff
or
describe
other
data
without
having,
like
other
things,
sort
of
hinder
your
interpretation.
So,
for
instance,
with
things
like
like
a
river
like,
you
may
want
to
simplify
a
representation
of
the
river
when
you're
like
say
like
measuring
like
river
flow
or
visualizing
river
flow,
like
maybe
too
much
detail,
and
too
much
squiggliness
can
sort
of
distract
your
eye
a
little
bit.
B
So
I
something
that
I
I
yeah,
I
sort
of
talked.
I
was
thinking
about
this
a
little
bit
like
when
giving
writing
up
these
examples
and
sort
of
where
I'm
at
with
these,
I
feel
like
everything
I
produce
out
of
our
and
qgis,
and
just
all
these,
like
I've,
tried
a
lot
of
different
tools
and
everything
still
feels
kind
of
technical,
and
I
think
it's
because
I
still
need
better
graphic
skills
like
it.
It
still
I
sort
of
want
to
go.
B
I'm
exploring
these
things,
because
I
love
drawing
maps
so
much
and
sort
of
doing
this
for
like
out
of
personal
pleasure,
but
I
feel
like
I'm
really
limited
by
and
I
want
to
see
if
I
can
do
this
with
software,
but
I
feel
like
I'm
a
little
bit
limited
by
yeah
a
little
bit
limited
by
arc,
but
something
that
a
lot
of
people
do
is
that
they
import
their
everything
yeah.
So
they
can.
B
You
can
import
or
export
stuff
that
you
generate
an
art
into
vector
formats,
so
you
can
open
this
up
in
illustrator
or
other
image,
editing
software,
like
even
the
example
that,
with
the
small
multiples,
I
showed
you,
I'm
pretty
sure
had
like
this
was
generated
in
r,
but
sorry
I'll
get
rid
of
this.
This
is
generated
in
r,
but
I'm
pretty
sure
there
was
some
adobe
illustrated
illustrator
applied
to
this
just
make
it
look
a
little
bit
more
polished,
but
in
terms
of.
B
Yeah,
if
you
had
wanted
a
spiral,
actually
I
think
a
spiral
would
be
possible
because
it's
it's
just
it's
sorry.
I
should
not
use
the
word
just,
but
it's
placing
things
on
a
spiral
is
math,
so
that
would
actually
be
a
really
nice
use
for
something
like
r,
because
you
kind
of
have
you
can
do
this
using
you
can
plot
like
things
in
a
center,
at
least
I've
done.
I've
done
similar
things
in
d3,
but
I
haven't
done
this
in
r
before
daniel.
B
I
guess
just
really
not.
I
don't
know
if
there's
much
really
special
out
about
my
art
flow,
I
mean
I.
What
I
really
like
about
r
in
r
studio
is
that
I
downloaded
rstudio
and
I
opened
it
and,
like
my
code,
is
there
and
after
you
know,
years
of
messing
with
them
and
emacs.
I
really
like
just
kind
of
having
everything
work
in
one
window,
so
I
tend
to,
for
instance,
require
all
my
libraries
at
the
top.
B
B
Actually
no,
I
already
have
a
parameter
for
this,
something
where
I
really
like
closure,
but
that's
the
same
value.
What
am
I
doing.
B
Something
where
I
really
like
closure,
because
this
kind
of
gets
it
can
get
kind
of
messy
and
big
and
not
having
my
editor
of
choice
is
a
little
frustrating.
A
I
think
we've
actually
got
to
follow
on
to
that
fairly
nicely
from
andrew,
which
is
I
use
r
mostly,
and
there
are
ways
to
get
many
plots
together.
A
couple
of
examples.
B
A
With
closure
are,
there
are
those
kinds
of
things
easy
to
do
and
then,
following
up
with
that,
is
the
for
database.
What
what
is
best
closure
or
are
why
I
guess
which
have
you
found
easier
to
use
or
produce
visualizations
with.
B
I
I
still
very
much
prefer
our
for
just
creating
output.
I
have
you
or
I
guess
my
question
to
andrew.
Have
you
used
anything
or
any
sort
of
visualization
library
in
closure.
B
There
is
a
library
out
there.
I
just
I
forgot
the
name
of
it.
B
A
Yeah,
I
think,
there's
a
few
different
options
with
closure,
but
what
I
hear
in
there,
which
makes
me
sad,
is
that
r
still
has
some
advantages
over
closure,
at
least
for
what
you're
doing
so
sounds
like
we
have
some
work
to
do.
Daniel.
B
It's
yeah,
it's
it's
hard,
just
because
yeah
r
has
like
they
they've,
been
sort
of
pushing
this
open
source
community
and
like
it's
used
by
basically
non-programmers
and
scientists
for
years
and
years,
and
there's
also
just
kind
of
like
this,
like
the
package
management
system
is
also
like.
There's
like
a
strict
submission
guideline
and
yeah
people
are
using
this
for
papers.
So
it's
it's
hard
to
yeah
compete
with
that.
A
B
I
I
think
this
is
expectation.
I
think
it's
a
good
stopping
point.
I
had
a
bunch
of
slides
about
this
is
going
to
go
into
a
little
bit
about
raster
files,
but
I
think
just
focusing
on
one
thing
is:
is
enough.
A
Sure,
okay,
then
opening
it
up
to
to
our
audience.
You're
welcome
to
continue
asking
in
chat,
we'll,
definitely
keep
an
eye
on
there,
but
feel
free
to
come
off,
mute
and
give
us
some
real
human
interaction.
A
I
think
I
did
see
one
come
in.
Are
there
any
other
examples
out
there
for
cow
plot
maps,
apart
from
visualizing
u.s
states.
B
D
Sorry,
maybe
I
didn't
answer
ask
my
question:
well
the
map
that
you
showed
for
the
u.s
states
in
which
the
states
were
not
limited
by
their
size
but
their
locations.
That's
what
I
meant.
I
thought
that
was
the
cowboy.
B
Oh
no,
no,
I
think
that's
something
else.
I
yeah,
I
don't
know
how
those
were
generated.
D
Okay,
because
I
think
the
u.s
states
have
straight
boundaries,
so
that's
that's
much
more.
That
kind
of
a
map
is
much
more
amenable
to
america,
but
like
for
india,
the
the
provinces
and
the
boundaries
are
very,
you
know,
haphazard.
So
I
don't
know
what
are
the
other
examples
where
those
kinds
of
maps
were
used.
D
B
I
don't
know
off
the
top
of
my
head.
I
don't
know
if
anyone
else
in
the
in
the
chat
knows
I'm
just
googling.
A
C
B
There
is
there's
a
couple.
B
I
just
want
to
see
if
I
can
find
the
europe
one
just
to
yeah
like
because
I
tend
to
show
a
lot
of
us
examples
but
yeah
I
just
it's
been
done
for
other
large
land
masses
that
are
not
in
the
united
states,
any
other.
B
A
I
have
kind
of
a
off-topic
question
from
where
we
were
at
well
andrew
asks.
So
you
said
you're
a
web
developer
by
trade.
Is
that
right.
B
A
B
I
so
whenever
I
made,
I
don't
do
so
much
like
one-off
visualizations
anymore,
but
when
I
used
to
I
used
to
use
things
like,
I
was
playing
around
with
r
and
python
and
trying
to
figure
out
which
one
was
the
best,
because
it's
it's.
Oh
yeah.
Sorry,
let
me
rewind
when
you're
creating
visualizations.
B
There
is
like
an
exploration
step
where
you
need
to
just
kind
of
make
a
bunch
of
visualizations
really
quickly,
and
I
I
kind
of
think
that,
like
web
development
is
really
poor
for
like
exploring
data,
I
it
just
because,
like
my
job
is
not
just
like
my
my
job
isn't
just
like
putting
stuff
on
the
screen
and
making
it
performant
it's
or
sorry.
My
day,
job
is
yeah
putting
stuff
on
the
screen
and
putting
graphics
on
the
screen
and
making
it
performant
and
run
well
and
also
look
nice.
B
But
when
you're
exploring
data,
the
purpose
is
to
just
get
as
many
visualizations
out
and
explore
and
like
see
the
data
from
different
perspectives
and
r
and
ipython
or
like
the
ipython,
notebook
environment
and
even
excel
spreadsheets,
are
really
really
good
for
that
and
yeah.
Also,
I'm
and
truly
like.
Sometimes
I
am
really
like.
I
just
want
to
get
away
from
the
browser
a
little.
A
Bit
yeah:
I
can
understand
that
so
we
got
two
in,
but
I'm
gonna
throw
it
to
andrew's
question.
First,
our
enclosure
advantages
and
disadvantages
of
each
broad
question:
when
do
you?
A
B
I
think
they're
they're
different
tools.
I
mean
right
now,
like
they're
different
tools,
and
I
I
think
what
daniel
is
building
with
clojuster
is
like
a
really
awesome
way
to
try
to
bridge
the
two
because
yeah,
like
I
think
r,
is
a
very
well
designed
language
that
I
I
really
enjoy.
Oh
sorry,
clojure
is
a
very
well
designed
language.
I
really
really
like
writing
closure.
B
I
think,
and
r
has
like
a
lot
of
really
really
great
libraries
behind
it,
but
I
think
it's
a
little
bit
clankier
than
closure
and
in
terms
of
database,
the
really
good
thing
about
r
is
that
it
really
just
takes
a
couple
lines
of
code
to
get
something
on
the
screen,
whereas
I
yeah-
I
don't
know
I
there
are
libraries
out
there
to
generate
graphics
and
closure,
but
they're
just
not
as
well
developed.
B
You
can
get.
You
can
yeah
download
a
csv
from
the
internet
and
download
rstudio
and
get
the
stuff
into
a
graph
within
10
minutes.
Even
if
you're
just
copying
examples
from
the
internet,
I
think
that's
a
really
big
advantage
to
our
so
yeah
and
then
also
yeah
like
it's.
I
think
closure
is
also
a
harder
language
to
get
started
in,
but
something
I
really
enjoy
with
closure
is
being
able
to
think
functionally
and
being
able
to
filter.
B
I
think
about
filtering
data
really
quickly
and
just
pulling
out
answers
that
I
I
need.
For
instance,
I
was
actually
developing
a
coding
challenge
for
our
team
and
I
was
just
testing
it
out,
like
writing
everything
enclosure,
script
kind
of
just
and
then
I
I
found
out
that
yeah
like
it
was
actually
quite
easy
to
get
something
up
and
running.
If
you
just
like,
do
the
hard
work
of
you
know,
setting
up
an
html
page
and
everything,
it's
just.
B
It's
just
easy
to
pull
out
data
that
you
need
rambling
a
little
bit,
but
I'm
done
for
my
answer.
A
Still
on
mute,
I
thought
that
that
was
actually
kind
of
a
an
interesting
answer
from
what
I
heard.
What
I
heard
you
say
is
r
is
very
easy
to
put
something
on
the
screen
right
generate
these
visualizations,
but
closure
is
actually
easier
easier
to
explore
the
data
with
when
you're
like
wait
a
minute.
I
need
to
work
with
this
and
let
me
filter
it
out,
and
you
know
those
types
of
operations.
Is
that
a
fair
summary
of
your
statement.
B
My
my
personal
opinion
is
that
I
don't
know
if
anyone
a
lot
of
other
people
would
agree
with
me,
I'm
just
I'm
also
quite
fluent
in
closure,
more
so
than
r.
B
So,
maybe
that's
why
I
think
like
that,
like
I,
I
tend
to
always
forget
our
syntax,
no
matter
how
often
I
work
with
it,
even
though
I
don't
work
with
it
professionally,
I
always
forget
how
to
do
things
like
just
even
like
applying
a
function
to
a
column
like
I
just
keep
forgetting,
whereas
yeah
closure
just
has
this
really
elegant
consistence
in
text,
if
you
were
to
like,
if
you
need
to
apply
a
function
to
a
list,
it's
just
yeah,
it's
kind
of
built
for
that.
A
I
think
that
actually
might
even
lead
us
into
the
next
question
from
jesus,
which
is.
I
was
thinking
that
java
is
already
a
famous
language
in
the
gis
ecosystem.
Have
you
have
you
ever
thought
about
exploring
the
java
gis
libraries
and
using
a
java
interop?
A
B
I'm
pretty
sure
I
tried
it
yes,
but
I
don't
remember
much
from
it.
I
I
I
yeah.
B
It's
I
mean
like
yeah
I
so
I
I
do
this
stuff
for
a
lot
of
this
stuff
for
fun,
because
it's
like
a
passion
for
me.
So
if
something
is
like
kind
of
a
bit
of
a
roadblock,
I
I'm
kind
of
the
type
of
person
who,
like
professionally,
of
course,
like
you,
get
through
it,
but
then,
if
you're
trying
to
do
something
for
fun,
sometimes
I'm
just
like.
No,
I
I
don't
want
to
spend
three
months
trying
to
set
up
this
environment
or
sorry.
Three
months
is
exaggeration,
but
yeah.
I
just
want
to
get
something.
B
B
I
I
do
someone
mentioned
yeah,
like
native
closure
functions
on
data
frames,
sometimes
yeah.
Sometimes
I
I
really
yeah,
like
I
pretty
much
have
to
like
balance
two
languages
or
figure
out
two
languages
in
my
head
and
kind
of
do
this
translation.
But
it
would
be
nice
to
have
a
more
closure,
syntax
for
say,
like
yeah,
taking
20
rows
from
something
or
yeah,
taking
a
range
of
something
or
yeah,
applying
a
function
to
something.
A
B
Oh
yeah
zoom
mass,
a
chat
window,
so
I
was
highlighting
the
question
thought
I
was
being
clever,
but
unfortunately
that
didn't
work.
A
Feature
anybody
else.
A
Just
gonna
give
it
just
a
minute,
but
thank
you
for
giving
us
this
presentation.
I
personally
am
also
interested
in
maps
and
I
I
wouldn't
call
myself
any.
I
guess
I
don't
know
the
word
cartography
nerd,
but
they're
always
enjoyable
for
me
to
look
at
so
I
I
had
a
good
time,
and
I
know
it
was
a
little
bit
of
a
struggle
with
emacs.
I
blame
emacs
for
any
type
of
presentation
problem
and
the
the
demo
demons
that
always
get
you,
but
I
thought
it
was
great.
B
Thank
you.
I
I
put
a
lot
of
effort
into
the
this
sort
of
example
directory.
I
think
yeah
these
examples
kind
of
speak
a
little
bit
better,
because
I
I've
also
haven't
given
a
talk
over
zoom,
so
I
was
mentally
just
not
really
understanding
what
was
going
on.
Just
I'm
just
speaking
into
my
computer,
so
yeah.
A
Interesting
times
we
live
in
where
we
we
have
presentations.
A
That's
also
like
a
chat
application
where
you
expect
just
going
back
and
forth,
but
no,
you
did
great,
and
it
was
my
first
time
moderating
our
chats
here.
So.
B
Awesome
you
did
a
great
job,
moderating
and
yeah.
Thank
you
for.
Thank
you
for
all
your
questions.
Everyone.
A
Yeah,
well,
I
guess
I
didn't
see
anything
else
come
in
so
if
I
think
joanne
you're
you're
in
our
zulip
chat
right.
So
if
any
anybody
can
direct
any
other
questions
there
and
follow
joanne,
I
think
you
said
on
twitter.
If
you
want
to
throw
that
up
one
more
time.