►
From YouTube: ML.NET End-to-End: Wine Regression with Database Data
Description
The first in a series of videos that goes end-to-end with an ML.NET model. This video goes over creating and using data in a database to create a model to predict wine quality and saves the data into Azure Blob Storage.
End-to-end 2: Web API video - https://www.youtube.com/watch?v=FFvpEeWW7AQ
Code - https://github.com/jwood803/MLNet-EndToEnd/blob/master/WineRegressionModel/Program.cs
Data - https://www.kaggle.com/rajyellow46/wine-quality
Contact:
Twitter: https://twitter.com/JWood/
Blog: https://jonwood.co/
A
Hey
everyone
so
I've
gotten
a
few
requests
for
building
my
dinette
models
using
data
from
a
database.
So
with
that
I
thought.
I
would
do
kind
of
an
Indian
series
where
I
would
take
data
set
and
put
it
into
a
database
and
then
read
from
that
database
to
build
an
internet
model
and
then
I'll
take
that
model
build
an
API
on
top
of
it
and
then
I
would
also
build
a
website
to
read
from
the
API
to
do
predictions
on
the
model.
So
that's
what
this
is.
A
This
is
the
first
of
the
series
where
I
will
take
this
data
set
here:
the
wine
quality
and
then
I'll
put
into
a
database,
and
then
I'll
read
from
it
to
create
a
regression
model
and
to
read
from
the
database
I'm
gonna
use
as
your
sequel
database
and
then,
when
I
build
the
model
I'm
going
to
save
that
model
into
as
your
storage
as
your
blob
storage.
All
right.
A
So,
let's
take
a
quick
look
at
the
one
data
set
here
and
we've
got
a
few
properties
of
material
such
as
a
type
whether
it's
red
or
white,
one
different,
acidities,
sugars
of
it
that
our
label
is
going
to
be
the
quality.
It's
gonna
be
a
rating
from
1
to
10
and
that's
what
we're
going
to
use
to
create
our
model
on
to
pre-dates
is
kind
of
the
quality
of
them
of
the
wine
based
on
this
properties
that
we
get
it
right.
A
A
And
give
a
database
name,
let's
give
it
one
data
create
a
new
server
I'll
leave
it
in
south-central
configure
the
database
here
make
sure
it's
the
cheapest
for
me
that
I
need
I.
Don't
need
much
for
this,
since
it's
just
pretty
much
the
demo
data
here
so
I'll
go
all
the
way
down
to
the
to
the
cheapest.
Then
I'm
going
to
create
that
all
right,
the
sequel
database
got
deployed.
A
A
A
A
Alright,
so
the
first
thing
I
want
to
do
is
add
a
configuration
file
to
hold
those
condition
strings
that
I
mentioned
edie
new
item
switch
for
Jason
here
now,
just
maybe
config,
and
if
you
had
a
new
foul
like
this,
always
make
sure
and
go
to
the
properties
and
market
to
coffee.
I
do
copy
of
numerous
others
and
always
copy,
and
then
I'm
going
to
add.
First
is
my
connection.
String
from
the.
A
A
A
A
A
A
A
A
A
A
A
A
A
A
Alright-
and
we
need
to
connect
to
our
database
here
and
for
the
server
we'll
go
back
to
ours.
Your
portal,
you're
gonna
click
on
that
database
and
we
get
a
server
name
here.
We
just
copy
that
with
that
in
there-
and
this
is
gonna,
be
a
sequel
login
and
it's
gonna
be
that
same
username
and
password
that
we
use
when
we
created
a
database
and
server.
A
A
Said
now
we
need
to
write
to
create
our
wine
data
table
hearing
so
I
just
paste
this
in
and
it
pretty
much
just
has
all
the
fields
that
we
use
for
a
wine
data
and
I
edit
a
an
ID
primary
key
in
here
as
well,
and
we
want
to
create
this
on
that
wand.
Data
database
that
we
just
created
so
I
just
run
that.
A
You
know
get
you
know,
of
course,
there's
nothing
in
it.
Yet.
So,
let's
write
the
code
to
add
items
to
our
new
database
here
now,
Rebecca
in
visual
studio
here
and
I
have
already
have
my
insert
command
here,
but
what
I
want
to
do
first
is
that
since
I'm,
probably
over
running
this
a
few
times,
I
want
to
make
sure
at
least
for
my
purposes.
A
A
A
A
A
A
A
A
Never
do
it
on
this
one
we're
getting
a
two-string
but
we're
assigning
to
a
float,
but
we
can
use
this
parse
method
once
again
and
parse
that
data
out
to
what
we
need
so
instead
of
typing
and
all
that
I'll
just
paste
in
the
rest
of
these
values.
Here
now
we
should
have
our
data
from
the
database,
which
is
going
to
be
a
list
of
one
data.
A
A
A
All
right,
so
we
got
here
I
step
over
it
preview.
Here
we
can
look
at
the
schema
that
we
have
loaded
here,
look
at
each
of
the
rows
that
we
got
and
a
call
name,
and
then
the
value
is
here.
So
it's
a
good
way
to
have
a
little.
That's
me,
my
cassette
has
a
preview
of
the
data
that
you
can
get
into
it.
A
And
the
next
little
pin
to
that
is
the
trainer
that
we're
going
to
use
regression,
trainers
and
I'll.
Just
do
a
fast
treat
trainer
yeah
like
a
specified
or
label
call
name.
If
you
want
yeah
that
we
need
to
here,
because
our
label
has
been
quality
and
we
could
put
the
features
column
name
here,
which
is
features
who
don't
have
to
since
that
is
defaulted
to
features.
A
They
pass
in
our
training
set.
So
now
that
we
have
our
model.
What
I
want
to
do
next
with
is
to
save
this
model
to
them.
There's
your
blob
storage
container
that
we
used
earlier
first
before
I
can
save
to
blob
storage.
Let
me
go
back
to
my
zero
portal
here
to
the
blob
storage
account.
I
can
do
this,
George
explore
I,
see
I,
don't
think
I
have
any
containers.
No,
let's
add
a
container.
Will
click
and
I
just
gave
up
models,
set
it
to
a
container
very
good.
A
A
A
A
A
Good,
but
this
upload
method
is
async,
so
we
can
put
a
weight
on
it
and
we
need
to
put
a
sink
and
our
method
here.
If
we
try
to
run
this
now
and
gives
us
an
error
saying
it's
not
suitable,
so
what
we
have
to
do
to
fix,
that
is
in
our
project.
We
can
edit
it
and
then
put
in
line
version
in
a
property
group.