►
From YouTube: Webinar | Two Outstanding Research Methods
Description
Each year, the National Legislative Program Evaluation Society recognizes up to three offices for reports that demonstrate outstanding use of evaluation research methods. In 2016, the Louisiana Legislative Auditor and the Colorado Office of the State Auditor won Excellence in Research Methods awards. During this webinar, Louisiana and Colorado audit staff explain their award-winning methods.
Louisiana staffers describe their method to measure how lottery ticket sales respond to changes in the prize payout percentage and how this method can be applied to estimate how a change in policy will affect outcomes in a variety of settings.
Colorado staffers describe how they applied GIS-based methods to the local sales tax system and how this method can be used to identify, analyze and communicate patterns and relationships that have a spatial dimension.
A
B
You
welcome
everyone
to
today's
webinar.
My
name
is
Linda
Triplett
and
I.
Am
the
director
of
the
Mississippi
care
committee
performance
Accountability,
Office
I
will
be
moderating.
Today's
webinar
I'd
like
to
start
off
by
thanking
the
National
legislative
program,
evaluation,
Society
and
the
National
Conference
of
State
Legislators
for
sponsoring
this
webinar
NLP
es
is
the
NCSL
Association
for
legislative
staff
engaged
in
program,
evaluation
and
performance
auditing.
B
Each
year,
the
national
legislative
program,
evaluation
society
recognizes
up
to
three
offices
for
reports
that
demonstrate
outstanding
use
of
evaluation,
research
methods
in
2016,
the
Louisiana
legislative
auditor
and
the
Colorado
office
of
the
state
auditors,
both
on
and
lt
es
excellence
in
research
methods.
Awards
during
this
webinar
staff
in
Louisiana
and
Colorado
will
explain
their
award
winning
methods.
B
C
Hello,
my
name
is
Christina
Wilson
and,
as
Linda
said,
it
will
be
presenting
information
from
a
performance
audit.
We
did
in
2015
on
the
Louisiana
Lottery
Corporation.
We
determined
that
the
lottery
that
maximized
the
lottery
profits
to
the
state
for
k-12
education.
If
they
increase
surprises
for
scratch-off
games.
Our
analysis
was
based
on
modeling
them.
The
demand
for
scratch-off
lottery
tickets.
To
give
you
some
background,
House
concurrent
resolution
116
at
the
2014
regular
legislative
session,
requested
us
to
analyze.
C
Of
lotteries
operations
generally,
our
audit
topics
are
determined
internally
by
our
statewide
risk
assessments.
At
this
time
this
was
a
special
legislative
request.
Once
we
first
began
the
audit
we
went
in
and
we
started
reviewing
lottery
financial
statements,
I'm
upon
looking
at
that
information,
we
saw
that
lottery
had
some
increases
and
some
of
their
revenues
that
we
wanted
to
see
what
best
way
they
could
maximize
the
amount
of
money
that
they
provide
to
the
state.
C
So
after
meeting
with
lottery,
we
determined
what
their
plans
were
to
do
with
those
increased
profits,
so
we
met
with
add
to
discuss
ways
that
we
could
present
meaningful
information
to
the
legislature
and,
at
this
time,
I'll
turn
it
over
to
Adam.
Who
will
discuss
the
details
of
the
analysis
we
provided
sure.
D
Thank
you
Christina.
So
what
we
did?
We
we
learned
that
the
lottery
had
saved
a
good
amount
of
money,
millions
of
dollars
by
switching
to
a
different
vendor
for
some
of
their
their
computer
systems,
that
they
used
for
the
Powerball
and
Mega
Millions
type
games,
and
so
one
thing
they
could
have
done
with
that.
They
could
have
just
transferred
more
money
to
the
state,
but
instead
they
decided
to
reinvest
it
in
prizes
for
their
scratch-off
games,
and
so
the
question
is:
did
that
was
that
actually
good
use
of
money
and
the
logic
behind?
D
That
is
that
if
you
put
more
into
prizes,
if
players
are
getting
more
in
prizes,
when
they
I
a
scratch-off
ticket,
they'll
actually
they'll
actually
buy
more
tickets,
and
so
that
was
the
logic
behind,
and
the
mode
we
found
is
we'll
explain
later
in
the
port.
Is
that
yes,
as
a
matter
of
fact,
that
relationship
does
exist?
In
fact,
if
you
were
to
increase
prizes
from
where
they
are
now,
which
is
about
63
percent
of
revenues
to
70
percent
of
revenues,
the
state
could
receive
an
additional
3.8
million
dollars
each
year
in
recurring
revenue.
D
So
we'll
get
into
a
discussion
of
some
of
the
other
policy
considerations
that
that
we
had
to
account
for
in
making
those
recommendations.
So
in
analyzing
sales
we
the
first
step
in
understanding
the
relationship
between
sales
and
prizes,
is
to
get
data
on
sales
and
prizes.
So
we
we
got
this
giant
data
set.
It
was
about
10
gigabytes
and
it
had
for
every
retailer
in
the
state,
know
every
little
gas
station
and
convenience
store,
and
then
it
had
for
every
day
of
the
year
and
then
for
every
game.
D
It
had
an
entry
with
how
many
tickets
were
sold,
how
much
they
paid
in
prizes,
and
then
that
sort
of
thing-
and
so
that's
a
little
bit
finer
detail
than
we
needed,
especially
when
I
was
you
know
five
years
worth
of
it.
So
we
we
summarized
it
by
perish
by
game
and
by
month,
and
so
that
that
gave
us
a
bit
more
manageable
data
set
as
worth
noting
what
these
games
are
the
way
to
scratch
off
tickets
work.
You
know
you
buy
this
ticket
as
a
design
on
it
of
some
kind.
D
They
have
themes
to
them,
so
one
of
them
might
be.
You
know,
I
love
the
New,
Orleans
Saints
or
you
know
I.
It
might
be
a
monopoly,
themed
game
and
they
come
in.
They
come
in
different
price
points
at
1,
2,
3,
5
and
$10
price
points,
and
they
last
for
about
12
to
18
months
at
point.
You
know
that
they've
really
stopped
selling
people
get
sick
of
them
and
they
change
it
out
with
something
new,
and
so
we
also
obtain
data.
D
You
know,
we
know
that
there
are
other
things
I'll
discuss
with
just
a
minute.
The
kind
of
model
that
we
develop,
we
also
obtained
data
on
unemployment
rates,
average
wages,
gas
prices,
we
generated
some
seasonal
effects
variables
to
add
to
our
data
set,
and
we
also
accounted
for
the
number
of
retailers
per
capita
in
each
parish.
D
So
we
also
only
move
on
to
the
next
slide,
the
so
the
demand,
the
demand
model,
is
kind
of
the
key
thing
that
we're
after
we
looked
at
some
of
the
literature
on
lottery
demand
and
the
way
that
economists
tend
to
talk
about
demand
is
its
responding
to
an
effective
price.
You
know,
ordinarily,
we
talk
about.
You,
know
the
price
of
corn
and
the
demand
for
corn
what-have-you.
Well,
in
this
case
it's
the
price
of
a
lottery
ticket,
but
it's
not
so
much
that
how
much
you
pay
for
the
ticket.
D
The
idea
is
what
you're
really
buying
is
a
lottery
prize,
and
so,
if
you're
getting
50
cents
back
in
prizes
for
every
dollar
you
spend
in
tickets,
then
that
means
that
you,
the
effective
price,
how
much
you
have
to
pay
for
a
dollar
prizes
is
two
dollars.
Likewise,
I
said
before
we're
at
about
a
sixty
2.75%
prize
payout
percentage.
D
So
if
you're
getting
sixty
two
point,
seven
five
cents
on
the
dollar
for
every
dollar
you
spend,
then
that's
an
effective
price
of
about
a
dollar,
fifty
nine-
and
so
that's
that's
kind
of
the
key
variable
that
they
include.
Now
it
just
so
happens.
This
is
basically
the
reciprocal
of
the
the
prize
pad
percentage,
and
so
it's
pretty
natural
to
go
back
and
forth
between
those
two.
D
We
also
wanted
to
account
for
the
number
of
months
that
the
game
has
been
open,
a
lot
of
cases
if
the
game
is
bound
for
a
while,
then
people
are
just
getting
tired
of
it
and-
and
you
know
we
wanted
to
account
for
that.
We
also
wanted
to
account
for
the
seasonal
variation
people.
People's
demand
for
lottery
tickets
varies
throughout
the
year.
It
tends
to
be,
for
example,
lower
in
the
summer
it's
and
that
sort
of
thing.
So
we
want
to
account
for
that.
D
We
also
wanted
to
account
for
the
number
of
retailers
each
parish,
the
idea
being
that
if
the
games
are
more
accessible,
if
there's
more
higher
density
of
places,
where
you
can
go
to
participate,
you'll
buy
more
tickets
and,
like
I
said,
we
wanted
to
include
the
the
unemployment
and
wages.
Now.
Those
two
are
kind
of
kind
of
interesting
to
include
in
there,
because
it's
it's
kind
of
ambiguous
beforehand
without
really
looking
at
it,
which
way
those
are
going
to
affect
demand.
D
In
other
words,
if
you
have
high
unemployment
is
that
going
to
raise
lottery
demand
because
people
feel
like
they're
higher
on
their
luck
and
they're
going
to
be
looking
to
to
buy
lottery
tickets
or
if
it
raises
unemployment,
are
people
going
to
say
well,
I,
don't
have
a
job.
I
need
to
cut
back
first
thing
to
go
as
lottery
tickets,
it's
not
clear
same
thing
with
wages,
so
we
wanted
to
include
that
gas
prices
we
felt
before
it'd
be
pretty
obvious,
which
way
those
should
go
mean
it's.
You
know.
D
D
Then
you
know
probably
that's
going
to
affect
the
decision
to
buy
a
lottery
ticket
and
we
also
included
separate
intercepts
for
each
price
point,
because
if
we're
looking
at
sales
sales
per
capita,
a
ten
dollar
game
is
just
going
to
mechanically
sell
more
than
a
one
dollar
game,
because
you're
getting
ten
times
as
much
per
ticket,
and
we
also
all
mentioned
we,
we
thought
about
trying
to
include
the
Powerball
and
Mega
Millions
jackpot.
But
you
know
that
kind
of
thing.
D
It's
kind
of
noisy
measure
just
because
you
know
there's
multiple
Powerball
and
Mega
Millions
jackpot
in
any
given
month,
and
they
don't
really.
The
lottery
doesn't
really
control
those
so
and
not
only
that
but
they're
pretty
random.
So
we
were
able
to
exclude
them
and
they
they
weren't
significant
in
any
event,
and
they
weren't
really
the
focus
of
our
study
because
they
didn't
they
don't
really
like
I
said
they
don't
really
control
that.
D
So
after
doing
all
this
after
we
developed
this
whole
model-
and
we
came
up
with
a
theory
of
how
these
things
relate
to
each
other
we
get
to.
This
is
really
what
we're
after
is
the
prize
elasticity
of
demand
and
it's
a
kind
of
technical
term.
But
it's
very
important.
This
is
the
percentage
change
in
sales
for
a
percentage
change
in
the
prize
payout
percentage,
which
will
denote
P
and
you'll,
see
it
in
the
formula
below,
and
the
idea
is
if,
if
the
prize
payout
percentage
goes
up,
one
percent,
not
one
percentage
point.
D
So
in
other
words,
if
it
goes
up
from
50
percent
to
fifty
point
five
percent.
What's
going
to
be
the
corresponding
change
in
sales,
and
we
can
estimate
this
using
a
Poisson
regression-
and
this
is
very
similar
to
a
least-squares
regression-
that
a
lot
of
people
are
familiar
with.
The
difference
is
that
it
has
to
do
with
the
fact
that
we
want
to
estimate
this
elasticity
parameter
when
you're
estimating
an
elasticity,
as
opposed
to
just
measuring
you
know
if
we
were
just
measuring.
D
You
know
dollars
of
change
in
sales
for
a
percentage
point
change
in
the
price
path
percentage.
We
wouldn't,
we
wouldn't
transform
those
variables
into
natural
logarithms,
but
since
we're
trying
to
get
an
elasticity,
we
took
the
natural
logarithm
of
sales
and
the
natural
logarithm
of
the
prize
payout
percentage.
But
in
a
lot
of
cases
there
were
games
that
sold
zero
tickets
in
a
particular
parish
in
a
particular
month.
D
And
if
you
try
to
take
the
natural
logarithm
of
zero,
then
you
get
negative
infinity
and
you
have
to
you
have
to
lose
those
observations
which
we
don't
want
to
do,
because
the
fact
that
a
game
has
no
sales
is
relevant
information.
We
want
to
include
that
in
our
analysis,
so
we
we
used
a
Poisson
regression
and
that
can
handle
zeros
in
the
dependent
variable
of
much
more
gracefully
and
and
that
log
link
equation.
I
mentioned
I.
D
Just
because
that's
what
still
enables
us
to
interpret
the
coefficients
as
elasticity's,
and
so
if
the
prize
path
percentage
increases
from
p0
to
p1,
then
the
increase
in
sales
is
is
going
to
follow
this
formula.
It's
you
know:
the
new
sales
s1
just
equals
the
old
sales
at
zero
times
that
ratio
p1
over
p0
to
the
power
of
epsilon
and
that
epsilon,
as
I
said
before,
epsilon
is
just
the
coefficient
on
the
log
of
the
prize
pass
percentage
from
that
regression.
So
that's
that's
where
it
comes
from
so
and
then
I
can
turn
to
this.
D
So
some
of
the
other
things
I
mean.
We
included
this
information
here,
just
because
it
was
interesting.
You
can
see
gas
prices
going
up
by
1%
decreases
sales
by
0.4
percent.
Approximately
wages
is
kind
of
interesting
it.
You
know
in
economics
they
talk
about.
You
can
have
a
normal
good
and
a
normal
good.
You
buy
more
of
it
if
your
income
increases
or
you
can
have
an
inferior
good
and
an
inferior
good.
You
buy
less
of
it
when
your
income
increases-
and
you
know
you
can
kind
of
cautiously
interpret
this.
D
This
negative
sign
here
as
telling
you
that
when
wages
go
up,
people
are
buying
less
tickets.
It
kind
of
implies
that
lottery
tickets
are
an
inferior
good,
which
is
an
interesting
term,
not
that
there's
anything
inferior
about
it.
That's
just
the
technical
term
that
economists
use
to
describe
it,
and
so
that
I
should
add
all
these
significant.
All
these
coefficients
were
significant,
but
we
also,
we
also
found
that,
as
expected,
retailers
per
capita
a
1%
increase
in
retailers
per
capita.
D
You
get
about
a
point,
8
percent
increase
in
sales
which
which
makes
sense,
and
then
finally,
we
looked
at
the
unemployment
rate
and
we
ran
a
bunch
of
different
models.
You
notice
this
is
the
only
variable
that
we
estimated
separate
coefficients
for
based
on
each
each
particular
price
point,
and
you
know
we
use
different
different
selection
criteria.
The
cake
information
criterion,
the
Bayesian
information
experience
to
pick
the
best
model,
and
it
was
it-
was
pretty
consistently
telling
us
that
we
needed
to
give
separate,
intercepts
or
separate
slopes
for
each
of
these.
D
Each
of
these
unemployment
rates
by
price
point-
and
you
can
see
that
I
think
the
reason
fed
is
pretty
clear.
The
one
in
$2.00
games
people
buy
more
of
them
when
there's
higher
unemployment
and
the
three
five
and
ten
dollar
games,
people
buy
less
of
them.
When
there's
more
unimportant,
so
they're
they're,
pointing
in
qualitatively
different
directions,
I
think
it's
it's
interesting.
D
And
then
we
say
what
would
happen
to
the
lotteries
financial
performance
if
you
adjusted
that
prize
payout
percentage
so
sales,
it's
pretty
easy
to
calculate
that.
You
know
we
had
the
formula
on
the
previous
slide.
It's
just
the
new
prize
path
percentage
over
the
old
one,
^
Epsilon,
that's
epsilon
2.88,
as
you
might
recall,
and
then
you
just
multiply
that
times.
The
FY
14
sales
and
you
have
the
new
scenarios
and
the
new
sales
in
the
new
scenario.
D
D
So
this
is
an
example
of
the
model
that
we
used.
You
can
see-
and
this
is
just
a
recreation
of
one
of
the
lotteries
of
the
2014
capper.
Just
selecting
the
key
variables
that
we
needed
to
look
at
the
top
line.
Here
is
the
scratch-off
prize
pad
percentage,
and
you
can
see
that
we're
starting
at
where
we
are
now
60,
2.75
percent,
increasing
it
to
70
point
four.
One
percent
and
now
everything
else
follows
from
that:
the
scratch-off
sales.
You
might
wonder:
okay,
221,
that's
an
actual
221
million,
that's
the
actual
number.
D
Where
did
the
307
nine
come
from?
Well,
we
just
took
70
point
400,
1/60
2.75.
That
gives
you
about
one
point,
one
two:
you
take
that
to
the
power
of
two
point,
eight
eight-
and
that
gives
you
about
one
point.
Three
nine
so
multiply
one
point
three
nine
times
the
221
million
and
that's
where
you
get
three
hundred
seven,
three
hundred
seven
point:
nine
million
from
and
that's
that's
where
it
comes
from
the
prize.
Expense
is
just
the
prize
path
percentage
at
the
top
times.
The
three.
D
Oh
seven
point,
nine,
and
so
that's
how
we
calculated
the
other
things
are
just
a
just:
a
percentage
of
sales.
The
draw
style
games
are
staying,
the
same,
fixed
cost.
Staying
the
same,
we
accounted
for
the
change
in
unclaimed
prizes.
If
you're,
if
you're
selling
more
tickets,
then
more
tickets
are
going
to
go
unclaimed,
and
so
we
wanted
to
account
for
that.
They
can
use
unclaimed
prizes
to
offset
their
other
expenses,
and
so
that
gives
you
a
net.
D
A
net
increase
in
your
net
operating
income
of
four
point:
three
million
dollars
and
just
to
mention
I
know
before
we
mentioned
it
was
about
three
point:
eight
million
was
the
increase
that
we
cited
in
the
report
and
that
accounts
for
some
other
non-operating
income
components,
particularly
interest
earnings,
which
will
describe
in
a
subsequent
slide.
But
that's
why
that's?
Why
there's
a
slight
difference?
D
D
D
So
here
are
some
of
the
scenarios
that
we
considered.
We
have
our
baseline
scenario,
that's
where
we
are
now
35
percent
and
a
63%
scratch
off
price
path
percentage.
Obviously,
there's
no
change
in
proceeds
to
the
state,
but
then
option
one
is
what
would
it
take
to
get
up
to
70%
and
that's
where
we
we
say:
well,
you
know,
if
you
want
to
be
conservative,
assume,
maybe
there's
some
jump
in
cost,
because
you
don't
want
to
you.
Don't
want
to
have
something
locked
in.
That's
that's
unrealistic.
D
And
so
what
we
wanted
to
show
is
okay.
Well,
if
you
took
away
that
fund
and
they
had
to
balance
their
budget
and
if
they
had
to
do
so
with
some
some
wiggle
room
in
case
they
had.
You
know
a
significant
drop
in
online,
the
the
draw
style
ticket
sales
or
what
have
you?
You
really
need
to
lower
that
scratch
off
price
pad
percentage
to
51%
and
if
that,
if
they
had
to
do
that,
the
state
would
see
a
16
million
dollar
decrease
in
proceeds.
D
So
and
ultimately,
you
know
it's
interesting
to
note
that
they
did
choose
option
two.
They
took
that
fun,
but
they
did
not
adjust
the
mandate
and
you
know
I
think
the
lottery
was
able
to
absorb
that,
because
in
the
subsequent
fiscal
year
they
had
record-breaking
Powerball
sales,
I
think
it
was.
You
know
in
national
news
how
high
the
Powerball
was
and
they
were
able
to
build
up
their
reserve
somewhat,
but
it
is
still
a
situation
that
they're
vulnerable
to
and
it's
something
that
we
do
plan
to
monitor
going
forward.
D
So,
in
conclusion,
we
talked
about
how
we
we
found
out.
They
were
increasing
prizes.
We
wanted
to
know
what
was
it
doing?
What
was
the
effect
of
it?
We
study
the
literature
developed
a
model.
We
estimated
that
model
to
see
what
the
relationships
between
the
different
variables
work
particularly
sales
and
prizes,
and
we
use
that
model
to
study
how
the
lotteries
financial
performance
would
change.
D
If
you
adjusted
prices
and
then
we
put
all
that
into
some
simulations
and
looked
at
it,
how
you
would
need
to
adjust
the
state's
policy
on
mandatory
transfers
to
the
state
from
the
lottery
in
order
to
realize
those
goals.
So
in
conclusion,
that's
that's
about
all
for
our
slides
and
with
that
all
I'll
hand
it
back
to
Linda.
Thank
you
very
much.
D
B
You
ed
and
Christina
I
would
like
to
remind
everyone
that
you
can
enter
questions
in
the
chat
box
at
any
time
out.
Very
good
I
see
another
question:
okay.
Next
our
speakers
from
Colorado
who
will
be
discussing
their
award-winning
performance
of
the
Department
of
Revenue
administration,
of
local
sales,
tax,
welcome,
Chris
and
Torrey
Thank.
E
We're
going
to
talk
to
you
today
about
our
audit
of
the
local
sales,
tax
collection
and
distribution
here
in
Colorado,
and
one
of
the
things
that
we
want
to
highlight
as
we
go
through.
This
is
the
way
that
we
use
a
geographical
geographic
information
system
tool
or
GIS
tool
to
use
a
mapping
for
in
our
analysis,
this
is
something
that's
new
to
us
here
in
Colorado,
and
so
we
wanted
to
just
sort
of
spur
your
imagination
for
ways
in
which
you
might
be
able
to
use
GIS
analysis
in
in
an
audit
or
a
program
evaluation.
E
Basically,
it's
because
there
are
some
questions
that
really
require
a
spatial
thinking
for
an
answer,
and
so,
for
example,
maps
can
help
us
see
and
analyze
concentrations
of
information,
for
example,
population
density.
It
can
help
you
feed
distributions
of
items
within
a
geographical
space,
for
example,
which
areas
of
your
state
have
more
hospitals.
It
can
also
help
you
analyze,
distance
relations
between
things,
for
example,
here
in
Colorado.
This
is
an
important
question:
are
there
any
marijuana
shops
and,
if
so,
how
many
that
are
within
a
thousand
feet
of
a
school?
E
And
it
was
occurs
to
me
that
even
in
this
presentation
you
know
they
there,
there
could
have
been
an
opportunity
to
use
GIS
to
analyze
the
relationship
of
retailers
to
gasoline
prices
that
could
have
been
represented
spatially.
So
for
us,
though,
most
of
us
are
familiar
with
the
way
GIS
can
be
used
to
create
maps
and.
E
We
have
been
thinking
that
you
could
use
GIS
sort
of
at
the
tail
end
of
an
analysis
safe
once
you've
done
some
analysis
on
some
data
and
you
could
load
it
into
GIS
and
create
a
nice
map
for
your
report.
Also,
it
occurred
to
us
that,
well,
you
know
you
might
use
GIS
sort
of
to
do
the
analysis,
in
which
case
the
end
product
would
be
again
again
be
some
map
that
you're
going
to
export.
E
The
aha
moment
came
for
us,
or
at
least
for
me
anyway,
I
will
say
when
I
realized,
that
GIS
is
not
just
a
graphical
tool
that
you
can
use
to
create
beautiful
maps,
but
it's
also
a
database
in
itself,
and
so
as
such,
you
can
use
it
to
create
data
tables
that
you
can
then
analyze
outside
of
GIS.
So
in
other
words,
on
this
project,
what
you'll
see
is
a
way
in
which
we
sort
of
flipped
the
tools.
E
If
you
will
and
use
GIS
at
the
beginning
of
our
analysis
and
then
did
more
of
the
heavy
lifting
outside
of
GIS.
So,
just
to
get
a
little
bit
of
an
understanding
of
what
this
project
was
about.
I'm
going
to
talk
about
local
sales
taxes
in
Colorado,
it's
a
complex
system.
We
have
about
297
cities,
towns,
counties,
special
districts,
local
improvement,
districts
and
transportation
authorities
which
have
the
ability
to
enact
the
sales
tax
on
any
sales
to
a
customer.
E
On
behalf
of
most
of
the
local
governments
in
the
state,
there
are
some
large
cities,
for
example,
Denver
Fort,
Collins,
Boulder,
Colorado
Springs,
that
are
what
we
call
Home
Rule
cities
and
they
are
able
to
they've
opted
out
of
the
Colorado
Department
of
Revenue's
process
and
they
and
they
are
responsible,
therefore,
for
collecting
sales
taxes
directly
from
the
retailers
within
those
within
their
boundaries.
But
for
most
of
the
counties
and
smaller
municipalities,
the
Department
of
Revenue
collects
that
sales
tax
and
then
turns
around
and
distributes
it
back
to
those
jurisdictions.
E
This
is
a
map
just
to
give
you
an
idea
visually
of
what
we're
talking
about
in
terms
complexity
of
the
sales
tax
system.
This
is
the
suburbs
sort
of
north
of
Denver,
the
Denver
metro
area,
and
what
you
see
here
is
I,
wanted
to
just
draw
your
attention
to
this
area.
I'm,
going
to
see
if
I
can
get
the
cursor
to
come
up.
E
Yeah
I
think
this
is
a
an
area
known
as
Erie
the
so
this
light
gray
area
shows
the
the
boundaries
for
the
town
of
Erie,
which
is
able
to
collect
the
sales
tax.
What's
interesting
is
the
regional
transportation
district,
which
is
indicated
by
the
diagonal
lines,
uncover
some
parts
of
Erie,
but
not
all
of
it,
and
even
within
Erie
there's
this
little
white
spot
indicating
an
area
that
is
not
incorporated
within
the
town.
E
Nor
is
it
subject
to
RTD
taxes,
so
you
can
just
get
an
idea
that
you
know
if
you're
a
business
in
this
little
area
here
indicated
by
the
arrow
it
would
be.
It
might
be
difficult
for
you
to
know
whether
you're
responsible
for
collecting
sales
taxes
for
the
town
of
Erie
for
RTD
or
or
maybe
none
of
them.
Maybe
you
think
that
you're
outside
of
Erie
and
only
would
them
collect
sales
taxes
for
the
county.
E
So
the
audit
was
really
generated
by
a
legislative
request
in
which
the
question
we
received
was
our
local
governments
receiving
the
correct
amount
of
sales
tax
revenue
during
our
initial
research
in
planning
the
audit.
We
discovered
that,
despite
that
complexity,
that
I
just
showed
you,
the
Department
of
Revenue,
does
not
have
or
use
a
map
to
determine
which
sales
taxes
apply
to
a
business.
They
have
no
GIS
tool
that
they
use
oil.
E
We
found
out
that
occasionally
they
will
look
up
a
business
address
in
a
mapping
system,
but
they
do
not
have
a
regular
process
for
for
that
kind
of
analysis,
and
so
we
saw
that
as
a
as
a
major
risk
area.
So
with
that
risk
area
in
mind,
we
kind
of
refocused
the
question
a
bit
to
look
at.
How
do
we
know
that
businesses
are
collecting
the
right
amount
for
the
right
jurisdictions
based
on
the
actual
physical
location?
E
Now,
a
key
control
in
this,
which
is
the
control
that
we
really
focused
on
in
in
our
in
our
audit,
was,
is
the
registration
process.
That
is
the
process
by
which
a
business
applies
for
a
sales
tax
license,
and
the
Department
of
Revenue
then
assigns
it
a
location
code
which
is
actually
printed
on
the
sales
tax
license
itself
and
that
location
code
is
what
the
Department
of
Revenue
then
uses
within
its
system
to
determine
the
sales
taxes
that
it
should
be
collecting.
E
So
you
can
imagine
a
lookup
table
if
you
will
for
every
location
code
that
represents
a
particular
area
of
the
state.
It
will
be
connected
to
all
the
jurisdictions,
so
the
county,
the
special
district,
the
city.
If
there
is
one
the
metropolitan
district,
there
could
be
others
housing
authorities
that
they're
supposed
to
collect
for
so
why
use
GIF
again
during
this
process
of
registering
a
business
for
self
tax
license?
E
The
Department
of
Revenue
relies
heavily
on
information
supplied
by
the
business,
but
what
we
found
was
in
a
lot
of
cases
that
that
information
really
isn't
sufficient
to
determine
the
precise
location
in
relationship
for
those
squiggly
line.
Boundaries.
Zip
codes
in
particular,
are
imprecise,
so,
to
the
extent
that
they
were
relying
on
goes
to
determine
their
taxation.
That
would
that
would
be.
You
would
you
would
get
a
lot
of
error
that
way.
E
So
the
ultimate
goal
for
us
was
not
so
much
a
map
that
we
could
put
in
our
report.
Although
we
did
create
maps
for
the
report,
but
we
wanted
to
create
a
table
of
data
showing
for
each
business
what
their
actual
tax
liability
should
be
based
on.
You
know
where
they
are
and
then
that
and
then
we
could
compare
that
table
to
the
data
that
the
department
had,
or
you
know
the
whatever
the
lookup
table
was
showing
based
on
their
location
code
and
see
if
there's
a
discrepancy.
E
F
Thank
you
Chris
and
good
day
everybody.
I
one
takeaway
that
I
hope
you
all
get
from
this
presentation
or
our
presentation
as
well
as
GIS
is
a
very
complex
and
big
system.
You
can
do
a
lot
with,
but
for
those
Jas
users
out
there
you'll
see
that
when
we
get
to
some
of
it,
when
I
get
to
the
slides
on
what
we
did
in
the
analysis
side,
we
didn't
do
a
lot
of
super
heavy,
in-depth
GIS
work
just
so.
F
This
is
to
show
that
you
don't
need
to
be
a
jazz
expert
to
do
automated
analysis
and
I
will
give
a
shout
out
to
my
former
office
and
other
NLP
ES
participant
the
Montana
legislative
audit
division.
Working
there
I
learned
GIS,
we've
got
to
start
with
GIS
and
learn
it,
but
they're
not
a
context
and
was
able
to
use
it
in
these
other
audits
to
show
that
we
can
do
some
spatial
analysis
that
has
a
pretty
big
impact
and
relevant
findings
for
for
performance
audit
work.
F
So
just
some
terminology
and
I'll
insert
here
too,
that
we
use
the
edge
route
every
software
package
of
our
GIS,
so
mainly
arcmap
10.3.
So
this
terminology
is
mainly
related
to
that,
but
it
applies
a
lot
to
a
lot
of
other
GIS
system.
The
GIS
data
can
be
represented
in
two
broad
ways:
vector
or
raster
data
vector
data
can
be
points,
lines
or
polygons
and
points.
F
Think
of
just
like
the
name
suggests
buildings,
mountain
peaks,
fire
hydrants
things
that
are
point
on
our
map
lines
like
rivers
and
slice,
pads
things
there
are
continuous
points
or
series
of
points
on
our
map
and
polygons.
Polygons
are
the
shapes
with
a
shape
and
close
by
a
boundaries,
and
then
we
can
calculate
an
area
from
things
like
cities
and
lakes,
and
even
buildings
can
be
represented
in
either
point
form
or
polygon
form.
The
other
type
is
raster
data.
Raster
data
is
a
little
more
complex
and
I'll.
F
Give
you
a
visual
example
on
the
next
slide,
but
it's
essentially
think
of
really
detailed
Landsat
images
or
aerial
photography,
type
data
and
what
we
can
do
with
raster
data
is
divider
slice
that
up
into
equal
sized
pixels
or
cells
and
each
of
those
pixels
or
cells.
When
we
click
on
it
or
go
into
it,
can
tell
us
various
types
of
information
within
the
layer,
so
a
matrix
of
raster
data
might
have
information
on
elevation
or
soil
saturation
or
things
like
that.
That
could
be
useful
for
land
planning.
F
So
here's
kind
of
the
world
broken
out,
as
you
can
see
on
the
Left.
We
have
kind
of
just
an
image,
though
it's
kind
of
a
SimCity
type
diagram
of
the
city
and
waterways
and
forests,
and
we
can
explode
that
on
the
right
into
the
various
layers
that
make
up
that.
That
will
view
such
as
imagery
and
addresses
water
features.
Things
like
that.
So
keep
this
visual
in
mind,
because
it
is
relevant
to
this
taxation
example.
F
We,
instead
of
imagery
elevation
and
transportation,
we
can
think
of
States
counties,
cities
special
districts
and
that
green
on
the
lower
left
is
just
an
example
of
raster
data.
The
smaller
image
shows
kind
of
an
aerial
or
bird's-eye
view
of
maybe
a
farm
or
some
type
of
land
use,
and
we
can
zoom
in
and
then
again
each
of
those
cells
or
pixels.
We
can
click
on
and
that
color
might
tell
us
something
about
elevation
or
pitch
or
so
low
saturation
and
different
things,
water
or
what
type
of
land
it
is
some
more
terminology.
F
So
each
of
those
layers,
each
of
geographic
things
that
we
can
see
on
our
map
has
data
underneath
it,
which
is
represented
in
GIS
and
what's
called
an
attribute
table,
and
with
that
attribute
table
we
have
each
field
can
store.
Cific
type
of
data
field
is
also
a
column,
so
our
screenshot
here
is
actually
from
our
local
sales
tax
data,
so
each
of
those
columns
might
have
an
address
or
a
city
or
a
county
or
whatever
business
it
is,
and
each
record
or
row
and
how
it
relates
to
that.
F
So
this
is
the
table
that
underlies
our
data
visually.
So
what
do
we
need?
Well
to
do
this
analysis?
We
first
needed
those
tax
jurisdictions.
We
needed
all
the
tax
boundaries
in
the
state,
all
the
layers
out
there,
and
it's
important
to
note
too
that
since
this
had
never
been
done
before
in
our
state,
we
had
to
do
a
little
bit
of
legwork
to
get
all
that
data.
So
we
talked
to
various
state
agencies,
local
governments,
cities,
counties
special
districts
on
pulling
together
all
the
taxation
boundaries
that
we
could.
F
Some
of
them
were
able
to
get
for
free
online
through
city
and
county
websites.
Some
we
had
to
contact
cities
and
counties
directly
and
some
we
got
from
various
state
agencies,
but
again
many
as
we're
out
there
online,
so
your
states
may
be
similar
and
that
most
of
the
GIS
data
is
actually
out
there
that
government
entities
are
publishing
for
public
use.
So
we
pulled
all
these
together.
F
The
different
layers
for
the
different
taxing
districts,
as
you
can
see
there,
and
then
we
have
the
business
addresses,
which
we
got
work
with
the
Colorado
Department
of
Revenue
to
get
the
addresses
for
all
the
businesses
that
are
registered
with
sales
tax
licenses
to
collect
sales
taxes
in
the
state,
and
then
we
need
a
tool
to
be
able
to
take
those
addresses.
You
know
the
one
to
three
main
streets
in
our
data
set
and
have
them
show
us
in
our
map
as
a
point,
that's
something
that
we
can
analyze
again.
F
So
that's
where
we
got
to
geocoding.
We
were
able
to
toke
the
businesses
and
had
that
tool
so
that
we
can
bring
them
up
on
our
map
and
geocoding
is
just
that
taking
a
real-world
location,
a
place
anything
from
a
house
to
a
mile
marker
to
whatever
and
making
that
represent
on
our
map
in
the
form
of
an
X
and
y
coordinate.
So
what
do
we
do?
F
Well,
first,
we
got
all
those
layers
in
the
Taxation
think
it
back
before
to
that
visual
example,
all
those
different
layers
for
cities,
County
special
districts,
and
we
combined
all
of
them
using
the
union
tool
into
one
multifaceted
layer.
So,
instead
of
15
layers,
we
had
one
layer
with
15
columns,
15
attributes
that
were
each
a
different
taxing
district
within
geo.
Coded
again,
we
assign
an
x
and
y
coordinate
reach
those
businesses
that
we
had
on
our
data
set
and
those
appeared
in
our
map
as
points.
F
Finally,
we
use
the
intersect
tool,
so
we
did
is
we
say:
okay,
GIS
come
combine
the
two
combined
all
of
our
tax
layer
and
our
point
and
super
second.
So
we're
essentially
saying
for
every
point
on
our
map.
Let's
drill
down,
if
you
will
and
tell
us
every
single
taxing
attribute
that
that
point
touches,
and
then
we
know
from
relief
from
art
in
rjs
universe,
for
each
address
that
we
had
in
our
map
what
tax
jurisdictions
are
relevant
to
it.
What
did
it
touch?
What
was
it
subject
to
so?
F
Here's
the
graphic
representation
user
that
I
spoke
a
couple
of
slides
of
screenshots
from
our
GIS
application.
Here's
the
state
of
Colorado
with
all
64
of
our
counties.
You
can
see
the
City
and
County
of
Denver
there
and
kind
of
a
mauve
Coral
salmon
color
in
the
center
in
the
center
and
then
just
to
the
west,
you'll
see
Eagle
County
and
the
beige
color
and
Beaver
Creek
as
it's
one
of
our
resort
communities
up
in
the
the
mountain
we're
going
to
zoom
in
on
the
Beaver
Creek
area.
F
So
now
notice,
Beaver
Creek
is
at
the
bottom
of
this
slide,
and
so
that
green
circle.
That
means
what
this
means
is
it's
not
an
incorporated
community
Beaver
Creek,
which
you
hear
a
lot
when
people
talk
about
this
area
of
the
mountains
is
actually
not
a
town.
It's
just
a
Norman
corporated
community
a
census-designated
place.
However,
those
red
lines
are
interest
rates.
Ebony.
This
is
along
the
interstate
70
mountain
corridor.
F
It
is
just
south
of
an
actual
town
called
a
bond
that
is
emit
is
an
incorporated
community
and
by
that,
by
extension,
it
adds
the
ability
to
enact
a
local
sales
tax
notice
that,
in
this
area,
there's
also
a
metropolitan
district,
which
is
a
form
of
special
district,
namely
the
bachelor
Gulch
MDT
is
for
the
left
in
the
green
color
and
all
of
those
points
are
the
businesses
from
our
data
set
that
we
geo
coded.
So
you
can
see
and
I'll
turn
on
our
my
pointer
here
just
to
kind
of
direct
your
attention.
F
You
can
see
there's
these
various
pockets
again
that
aren't
incorporated
or
0.
These
points
like
this
one
that
are
right
on
a
boundary.
So
it's
really
to
make
sure
we
have
accurate
representation
of
where
these
businesses
are
in
relation
to
the
incorporated
on
corporate
areas
of
a
community
or
taxation
district.
F
So
this
just
gives
a
graphic
representation
of
again
what
we
saw
and
then
what
we
did
at
this
point
was.
We
took
that
attribute
table
that
I
mentioned
before
that
had
all
of
the
results
of
combining
our
tax
districts
and
our
points
of
business
addresses
and
then
to
do
the
next
of
our
analysis.
We
exported
it
to
Microsoft,
Access
and
so
I'll
hand
it
back
to
Chris
and
he's
going
to
talk
a
little
bit
about
how
we
did
this
work
in
Access.
E
Thank
You
torrei,
so
the
results
of
our
analysis
in
GIS
presented
a
table
that
looks
something
like
this
on
the
left.
You
see
the
address
of
businesses
as
we
received
them
from
the
Department
of
Revenue
on
the
right.
You
see
the
columns
that
were
added
by
the
GIS
tool,
and
so
each
column
represents
a
different
type
of
taxing
jurisdiction
and,
what's
what's
great
about
this,
is
if,
if
a
if
a
business
is
in
an
unincorporated
area,
for
example,
here
I'm
going
to
turn
the
cursor
back
on
here,
you
see
this
three
nine
397
Prater
Lane.
E
It
says
it's
in
the
town
of
dazon,
according
to
Department
of
Revenue,
but
in
reality,
according
to
our
GIS
work,
you
see
that
this
has
a
null
value
in
the
city
fields,
which
means
that
it
isn't
actually
touching
any
or
it
isn't
touching
or
within
any
incorporated
cities
boundaries.
So
that
means,
even
though
it
has
an
Avon
address,
it's
not
actually
in
the
town
of
Avon,
but
it
is
in
Eagle
County
and
it
is
in
the
bachelor
Gulch
district.
So
it
should
be
collecting
taxes
for
those.
E
If
you
look
here,
though,
you
also
see
an
example,
another
example
down
here
row:
16,
here's,
a
here's,
a
city
that
had
an
Englewood
address
I
mean
a
business,
had
a
Englewood
address,
but
it's
actually
in
the
town
of
Centennial.
So
if
that
business
was
collecting
taxes
for
Englewood,
then
it
would
be
doing
so
improperly.
It
should
be
collecting
taxes
for
Centennial.
E
So
it
was
that
sort
of
discrepancy
that
we
needed
to
then
analyze
which
we
couldn't
do
in
GIS,
but
we
we
did
within
access
to
not
only
see
did
the
differences
between
the
reported
address,
but
also
those
woke,
those
lookup
codes
that
I
was
taught
those
location
codes
that
I
was
talking
about.
So
what
we
did
then
was
we
created
a
similar
table
of
based
on
the
loop,
those
location
codes
to
represent
all
of
the
taxing
jurisdictions
that
the
Department
of
Revenue
said
that
the
businesses
should
be
collecting
for
and
then
compared
it
to.
E
This
was
the
results
that
you
see
here,
and
the
results
of
that
was
that,
after
excluding
some
businesses
that
had
really
strange
addresses
like
corner
of
12th
and
main,
which
were
difficult,
you
know
virtually
impossible
to
map
with
the
tool
that
we
had.
We
really
needed
standardized
addresses.
We
were
able
to
nevertheless
map
100
3836
businesses,
which
is
virtually
the
whole
population
of
businesses
that
were
registered
to
collect
sales
taxes.
For
that
year
we
found
that
about
11,000
of
them
were
Mis
registered.
That
means
they,
they
were
showing
a
location
code
that
had
tax
implications.
E
That
we
showed
were
incorrect.
That's
about
11
percent
error
rate,
so,
as
I
said
before,
the
Department
of
Revenue
does
not
collect
sales
taxes
for
certain
cities.
So
when
you
exclude
the
Home
Rule
cities
that
collect
that
are
required
to
collect
for
them
selves,
the
error
was
about
3746
businesses
were
Mis
registered
for
jurisdictions
that
the
Department
of
Revenue
collects
South,
Texas
for
and
just
looking
at
those
3746,
not
the
total
11
thousand.
We
found
that.
E
E
So
you
see
the
numbers
they're
about
3.3
million
on
the
over
question
and
3.8
million
on
the
under
collection
side.
An
interesting
side
benefit
of
this
analysis.
That
we
did,
though,
was
that
we
found
that
there
were
a
few
quite
a
few
businesses
anyway
or
enough
to
be
concerned,
but
for
which
we
were
unable
to
sign,
assign
a
location
code,
and
the
reason
is
because
they
they
were
in
an
area
that
has
a
a
combination
of
taxing
districts,
for
which
the
Department
of
Revenue
did
not
have
a
location
code
established
for
that
particular
combination.
F
Thanks
good,
so
yes,
Chris
mentioned
when
we
tried
to
do
that:
reverse
lookup
as
Chris
mentioned,
and
say
all
right.
Well
for
all
these
ones
that
are
in
the
wrong
one
that
has
the
wrong
location
code.
What
should
their
location
code
be?
And
we
found
up
about
nine
areas
in
the
state,
and
this
is
an
exhibit
from
our
report
that
shows
two
of
them
where
the
combination
of
county
city,
special
district,
that
that
the
area
was
there
was
a
location
code
for
so
the
town
of
Watkins
on
the
determiner
of
a
con
over
there.
F
So
this
town
of
Watkins
down
here
is
actually
unincorporated
community
but
notice
how
it's
in
that
RTD
is
in
that
hashed
area.
But
what
we
found
was
there's
only
one
location
code
for
Watkins
and
did
not
have
the
RTD
district
or
regional
transportation
district,
so
that
was
an
example
of
under
collecting.
If
you
were
business
in
that
area,
you
were
not
being
required
to
collect
reads:
RTD
sales
tax.
Conversely,
these
areas
of
Aurora,
the
town
of
Aurora
in
Adams
County,
the
only
location
codes
that
are
incorporated
Aurora
in
Adams,
County
included
RTD.
F
So
these
areas,
however,
we
found
that
there
was
no
such
code
for
Aurora
in
an
accounting
book
out
RTD,
so
these
businesses
were
over
collecting,
they
were
collecting
RTD
when
they
didn't
need
to
literature
shows,
as
Chris
mentioned,
have
an
ancillary
benefit.
We
had
of
doing
this
analysis
as
we
found
there's
up
also
holes
in
the
system,
so
the
businesses
are
collecting
the
right
amounts.
So
what
is
that?
F
That's
our
presentation
on
how
we
use
GIS
in
a
performance
audit
context,
and
oh
yes
and
I'll-
mention
real
quick-
that
we
did
issue
recommendations
to
Department
revenue.
They
agree
with
all
of
them
and
they
are
working
to
implement
a
GIS
system
to
do
enhance
the
registration
of
businesses
for
sales
tax
licenses.
So
with
that,
thank
you
very
much.
B
Thank
you
Chris
and
Tori.
We
appreciate
all
of
it
excellent
presentations
that
we
have
heard
today.
We
will
now
move
into
our
question
answer
period
and
I
see
we
don't
have
a
great
deal
of
time
and
thank
everyone
for
the
questions
they
have
submitted.
I
will
direct
the
questions
to
our
panel,
but
if,
if
we're
not
able
to
answer
them
all
in
the
timeframe
that
we
have,
the
panelists
have
agreed
to
answer
them
offline
at
after
the
webinar,
okay.
B
D
Thanks
thanks
mr.
Fugate
for
your
question
that
we
actually,
we
showed
them
our
analysis
and
it's
comported
with
what
they
had
been
saying
for
a
long
time,
which
is
if
they
really
should
be
at
a
higher
price
paid
percentage,
so
that
I
think
they
were
supportive
of
the
findings.
They
use
a
similar
model
from
a
vendor.
They
like
to
talk
about
that.
You
know
the
fact
that
it
has
64
variables
like
it's.
You
know
Heinz
57,
flavors
or
whatever
it
is
it's
it's
kind
of
their
their
their
their
trademark
thing
I,
think
they're.
D
Their
biggest
concern
was
that
the
it
was
the
policy
recommendation
surrounding
the
mandate
in
the
fund
and
I
think
they
foresaw
that
it
was
just
going
to
result
in
the
loss
of
reserves
and
that
nothing
was
going
to
happen
to
the
mandate
and
that's
ultimately,
what
happened
so
I
think
they
were
supportive
of
the
analysis
that
we
did
maybe
a
little
bit
concerned
about
putting
the
results
up
out
there
and
the
the
political
upshot
of
doing
so
so
I
hope
that
answers
the
question.
Okay,.
B
E
Thank
You
Linda
I'll
take
that
one
Thank
You
mr.
Jutras
for
that
question.
If
a
business
has
multiple
addresses
in
Colorado,
they
were
required
to
have
separate
sales
tax
license
for
each
location,
so
you
know,
and
so
they
would
be
reporting
sales
for
those
locations
and
interesting
thing.
That's
related,
though,
is
that
if
a
business
has
sort
of
no
physical
location,
maybe
has
or
maybe
has
delivery
trucks
that
travels
around
the
state
that
pose
an
interesting
problem
that
was
a
whole
separate
set
of
businesses
that
we
were
not
looking
at.
E
We
were
not
looking
at
those
types
of
businesses
because
there's
not
a
whole
lot.
We
can
do
to
determine
without
getting
their
delivery
receipts
to
identify
the
exact
homes
that
they
were
visiting
or
whatnot,
however,
that
we
did
discuss
that
in
our
report,
the
complexity
of
that
and
so
for
those
businesses
that
do
that
they
have
to
pretty
much
rely
on
the
self
reporting
of
the
businesses
in
terms
of
where
the
sale
took
place.
B
D
I
think
to
do
that,
we
really
would
have
needed
to
look
at
it
on
a
daily
basis
and
look
at
the
daily
Powerball
and,
and
that
would
have
been
I,
think
it's
a
very
possible
analysis.
I
think
it's
certainly
something
that's
worth
doing.
I
think
the
challenge
is
that
we
would
just
need
a
larger
data
set.
It
also
really
wouldn't
be
something
that
we
could
recommend
to
our
lottery,
since
they
really
only
have
one
vote,
but
it
might
be
something
very
interesting
for
all
of
them.
D
You,
a
multi-state
lottery,
Association
states,
since
you
know,
they've
been
doing
all
this
work
to
try
to
increase
Powerball
jackpots
and
increase
Powerball
sales
in
it.
It
might
be
interesting
enough.
That's
just
cannibalizing
one
of
their
other
business
lines.
So
thank
you
very
much
for
your
question.
I
hope,
I
answered
it.
B
F
This
is
Tori
thanks
for
the
question:
Kaitlyn
Thank,
You
Linda
down
the
free
straightfoward,
the
first
one
G
is,
has
a
geocoding
tool
within
it,
but
the
number
of
addresses
we
were
wanting
to
geocode.
It
was
going
to
actually
gonna
cost
a
fair
amount,
so
part
of
the
work
we
did.
We
ended
up
in
researching
layers.
We
were
working
with
our
state's
office
of
information
technology,
Oh
IT,
which
is
an
executive
right
agency
and
they
have
a
GIS
unit
within
it
within
them
and
they
have
developed
our
own
geocoding
tools.
F
We
have
an
in-state
geocoding
tool
that
made
life
quite
nice
and
they
rely
on
census
data
Tiger
data,
if
you
familiar
with
those
data,
sets
to
to
develop
an
in-house
geocoding
tool.
Okay,
on
your
other
question:
yes,
that's
kind
of
an
always
effecting
problem
in
the
GIS
world.
We
did.
We
had
layers
that
want
to
different
cornices
and
for
our
listeners
out
there
think
back
to
grade
school.
Think
of
your
wall,
maps
of
the
world
and
think
of
the
different
types
number.
F
The
the
problem
is
converting
the
round
world
into
a
square
or
rectangular
interface,
and
we
have
to
stretch
it
and
tweak
it
in
and
make
it
fit
that
shape
so
think
of
the
maps
where
Greenland
is
gigantic
and
in
green,
that
is
already
large
or
Canada
is
gigantic.
That's
one
type
of
projection
or
the
type
where
so
yeah
the
various
out
there.
F
B
C
F
Karen
and
yeah
thanks
for
that,
it
was
being
paid
when
we
say
over
collecting
that's
what
they
were
doing.
They
were
actually
collecting
that
text
and
remitting
it
to
the
Department
of
Revenue,
so
that
the
question
of
pocketing
is
more
of
the
Department
of
Revenue's
auditors.
That
would
go
out
and
audit
businesses
to
make
sure
they
were
submitting
the
correct
amounts
of
sales
tax.
But
what
we
looked
at
was
the
businesses
were
actually
turning
in
that
money
to
the
Department
of
Revenue.
B
Okay,
well,
thank
you,
we're
right
up
at
an
hour,
so
thank
you
again
so
much
ed,
Kristina,
Chris
and
Tory,
and
thanks
to
all
of
you
for
joining
us
today
again,
this
webinar
webinar
has
been
recorded
and
will
be
made
available
on
the
NPSL
website
in
the
next
couple
of
weeks.
Thank
you
again
for
your
participation.
This
webinar
is
now
concluded.