►
From YouTube: Augur - Sean Goggins
Description
Speaker: Sean Goggins
Organization: University of Missouri
Risk and Value metrics are initiating a process of aggregating CHAOSS metrics into new forms. Augur's integrated git, Issue tracker and code evaluation tools to address these needs in a new, consolidated data model, supported by GSoC students and ongoing development. This presentation will review Augur in the context of Risk and Value metric implementations.
A
So
augur
has
been
defined
for
a
long
time
around
the
idea
of
being
able
to
do
comparisons,
download
your
data,
look
at
any
repo
group
on
github
or
not
see
the
issues,
the
comments
things
like
that,
but
all
in
one
place.
So
some
of
you
are
familiar
with
what
we've
done
before
I
want
to
tell
you
a
little
bit
about
the
things
that
you
cannot
see
here
right
now.
A
The
first
is
that
we
now
gather
all
data
from
all
sources
into
an
organized
relationally
structured
database,
so
that
the
integrity
of
the
data
that
we're
collecting
is
guaranteed
not
just
by
the
integrity
of
a
normal
database,
but
by
tests
that
we
have
that
ensure
that
that
have
that
things
happen
correctly.
The
second
we
have
a
distributed
worker
architecture-
that's
easy
to
extend
information
to.
So
if
you
have
something
that
you
want
to
collect,
we
have
several
models.
A
I
think
we
have
eight
workers
that
are
working
right
now
that
Wilco
out
and
you
can
define
where
you
want
to
get
the
data
from
what
api's
you
want
to
call
how
to
collect
it,
and
then
you
can
build
an
API
for
it,
and
so
this
is
the
the
front
end
of
augur,
which
is
the
pretty
part,
but
there's
a
lot
of
technology
behind
it.
That
I
think
is
really
interesting
and
important
to
understand
so
I'm
going
to
highlight
a
few
things.
The
first
thing
I'm
going
to
highlight
is
insights.
A
So,
if
we're
able
to
collect
information
on
over
60
different
specific
metrics,
so
we
have
metrics
that
are
included
in
chaos
and
metrics
that
we've
conceived
while
working
with
organizations
that
are
part
of
chaos
and
haven't
yet
gone
through
the
process
of
making
into
chaos
metrics.
If
we
can
conceive
of
these
60
metrics
and
the
question
what's
interesting
or
what's
changed
is
often
what's
asked.
How
do
we
tell
you
that?
How
do
we
let
you
have
an
easy
view
of?
A
So
if
I
click
on
this
I'm
gonna
see
what
is
what
is
interesting
about
risk,
so
it
would
have
told
me
I,
guess,
I
didn't,
show
you
or
didn't
read
to
you
back
here
is
this
repository?
Has
a
sharp
decrease
in
commit
count
within
the
past
year?
Basically,
so
right
now
we're
looking
at
timeframes
for
the
past
year,
but
you
can
set
that
for
90
days.
You
can
set
that
for
a
week
you
can
look
for
anomalies
over
whatever
period
you
want.
A
The
top-10
committers
are
listed
numerically,
and
then
we
can
see
information
about
lines
of
code
added
over
time,
the
total
ratio
of
lines
of
code.
If
we
had
organization
information,
we
could
show
that
and
then
this
looks
at
lines
of
code
added
the
top
10
authors
over
the
life
of
the
project.
So
you
can
see
who's
committing.
What
deleting
adding
where,
where
let
me
stop
sucking
the
air
out
of
the
room
and
ask
if
anyone
has
any
questions
about
what
I
have
described
so
far.
B
A
So
I
think
I
understand
the
question
its.
If
we're
looking
at
repose,
how
do
we
compare
repose
that
have
a
very
low
volume
to
begin
with
and
repose
that
have
a
very
high
volume
to
begin
with,
so
we
are
looking
statistically
at
the
absolute
difference
in
the
context
of
that
repo,
so
something
equivalent
to
a
z-score.
A
So
within
the
repo
it
isn't
if
I
have,
if
one
repo
has
an
average
of
30
commits
a
week
and
another
repo
has
an
average
of
30,000
commits
a
week,
a
significant
change
in
the
30
repos
is
going
to
show
up
just
as
quickly
as
a
significant
change
in
the
30,000
repos
and
vice
versa.
So
we're
looking
at
the
repos
in
the
context
of
themselves.
We
can
also
calculate
across
the
repository
group
to
see
if
the
group
as
a
whole
is
is
doing
things,
and
we
one
of
the
things
that
I
think
is
really
interesting.
A
A
C
A
Is
and
it
used
to
be
extremely
hard,
I
think
it's
a
lot
easier
now,
Gabe
and
Carter,
who
are
developers
for
August
and
can
tell
me
tell
you
a
lot
more
about
the
details.
We
made
a
couple
critical
changes
over
the
summer
with
our
google
Summer
of
Code
students.
One
is
we
changed
all
the
front
end
work
from
typescript
or
from
JavaScript
to
typescript.
That
makes
it
more
definitive
and
it
also
requires
a
lot
less
code.
So
it's
a
lot
easier.
We've
also
included
those
steps
that
you
use
to
add
a
view.
A
Visualization
I
think
are
more
clearly
understandable
than
they
were
before
so
I
think
you
do
have
to
know
a
little
bit
about
view
and
view
Jas
and
vcli
to
add
a
visualization
and,
as
you
know,
those
you
know,
some
of
the
lot
of
people
in
this
room
are
not
front-end
technology
people,
but
we
do
have
templates
that
we,
you
know
each
metric,
that's
here
you
can
copy,
and
these
metrics
come
out
of.
What's
puts
them.
B
A
So
in
my
in
my
experience,
organizational
affiliation
is
a
difficult.
It's
difficult.
Typically,
it's
managed
their
tooth
there.
Two
things,
one
is
an
organization
within
its
own.
Boundaries
can
manage
those
affiliations
in
that
list
without
any
implications
for
gdpr
or
the
other
privacy
laws
that
are
are
coming
into
play.
So
that's
thing
one,
and
so
they
can
maintain
those
Maps
and
some
companies
that
we
work
with
have
started
to
do
that.
A
The
second
thing
is
that
we
have
a
worker
or
a
sort
of
a
model
that
we
use
now
that
we
we
go
through
all
the
git
repositories
collect
all
the
emails
and
then
search
various
web
api
is
to
identify
organizational
affiliations
or
github
ids
for
those
individuals.
So
to
the
extent
they
keep
information
up-to-date
on
other
web
properties,
we
can
make
probable
guesses
about
that.
But
those
are
not
that's
not
good.
Enough,
typically
I
mean
I,
think
there's
you're,
always
if
you're
really
wanting
to
manage
an
organizational
affiliation.
A
A
You
know
I
think,
for
certain
reasons,
I'm
afraid
to
share
certain
information
about
myself,
online
and
I.
Think
a
lot
of
people
are
the
same,
but
if
we
offer
them
a
secure
way
to
do
it,
that's
in
a
federated
distributed
encrypted
form,
like
you
know,
hyper
ledger
in
D.
You
know
I
think
that's
a
contribution
that
chaos
project
could
potentially
make.
A
A
And
if
I
just
type
that
slack
app
and
the
word
insights
it's
going
to
and
we
can
have
it
triggered
by
other
words
and
we're
also
going
to
push
it,
but
it's
basically
able
to
send
us
now.
Remember
I
showed
you
this
this
top
insights
page
over
here.
So
whatever
the
top
insights
page
is
for
your
projects.
If
you
have
a
new
insight
that
surfaces
on
a
day,
we
have
the
ability
for
you
to
add
a
slack
notifier
into
a
slack
channel
and
it'll
just
send
you.
It
and
it'll.
Send
you.
However,
many
slack
notifications.
A
A
A
I
previously
installed
the
app
in
the
site
so
the
first
time
you
do
it
in
the
channel
it'll
install
it
in
your
slack
site
in
the
second
time,
you're
just
subscribing
a
channel
to
it,
mm-hmm
exactly
and
I'm.
Looking
for
a
unthink
endure
as
ad
of
support
there
was.
It
was
great
Jonah,
our
developer,
that
did
this
sent
a
bunch
of
these
yesterday
that
had
like
too
many
replies.
A
So
what
I
think?
What's
there's
a
small-knit
in
the
design
where
you
all
know
how
you
could
be
overwhelmed
by
slack
like
if
a
notifier
sent
you
30
messages
in
one
day
that
would
be
super
annoying,
so
the
notifier
will
send
you
one
message
and
then
thread
like
if
you
have
30
interesting
notifications,
because
you're
monitoring,
5000
repositories,
it's
going
to
send
them
all
in
one
post
and
then
all
the
individual
details
under
a
thread
so
that
it's
not
a
nuisance
for
people
that
are
examining
large
numbers
of
repositories.
B
A
C
A
Now
we're
defining
what
an
anomaly
is
that,
if
it's,
if
it's
like,
if
we
take
a
three
like
a
one-year
average,
we're
gonna
say
that
and
you're
interested
in
things
that
happen
in
the
last
month.
We're
gonna
notify
you
of
things
from
the
last
month
that
that
are
anomalous
in
the
context
of
the
previous
year,
yeah
that
Gabe,
you
want
to
add
any
details
to
that.
B
A
We
also
have
a
risk
page
which
matches
Matt's
gonna
talk
about,
so
I
won't
steal
that
you
can
also
see
a
complete
list
of
the
repos
there's
2200
repos
that
are
listed
in
this
particular
example.
You
can
sort
them
by
the
total
number
of
issues
identified.
Obviously,
some
repositories
have
zero
issues
because
they
don't
use
github
issues.
A
We
haven't
put
in
a
description
because
I'm
lazy
you'll
see
the
total
commits
accounts.
The
total
issue
counts,
I
can
sort
by.
However,
many
of
these
I
want
to
do.
I
can
see
a
list
of
the
repo
groups
and
then
go
under
here
to
see
which
repos
are
under
each
repo
group.
These
are
the
ones
that
were
we're
currently
tracking.
A
All
right,
yeah
I
do
so
I
just
want
to
go
through
the
like
active
guessing
here,
because
some
of
them
have
no
issues
whatsoever
and
the
comparisons
that
we're
showing
right
now
are
issues.
So
here's
a
case
where
we
can
see
two
different
repositories.
The
closed
issues
per
week
and
this
the
opened
issues
per
week
over
time
and
how
they've
changed
and
you
can
see
each
repository
is
a
different
color
and
you
can
compare.
Is
it
up
to
like
eight
different
repositories
that
you
can
compare?
A
A
We've
also
continued
the
the
thing
that
we
had
in
the
older
version
of
augur,
which
is
we
can
download
as
an
SVG
a
PNG
view,
the
source
or
open
and
a
V
get
a
deterrent
fj's
on
download
already
implemented
as
well,
which
these
are
long-standing
features
in
augur
that
let
people
take
vis
from
the
auger
and
just
put
it
in
the
PowerPoint
or
data
and
share
it
with
other
people.
So
that's
got
to
be
five
minutes
right.
Okay,
any
last
questions.
A
So
the
frankly
historically,
we've
been
gluing
together
a
lot
of
different
work
and
it's
been
not
a
service
model
by
some
kind
of
intricate
design,
we're
academics.
Now,
if
there's
no
business
model,
yet
you
can
upload
me
in
the
last
over
the
summer.
I
think
we've
got
it
pretty
close
to
the
point
where
we
have
an
install
script:
a
startup
script
to
shut
down
script,
things
that
let
you
just
kind
of
load
you
know
clone
it
and
load
it.
A
The
most
there's
a
few
steps,
I
mean
there's
a
very
detailed
set
of
instructions,
so
anybody
can
technically
download
it
and
install
it
and
we're
making
it
easier
and
easier
for
people
to
do
that
everyday,
but
there's
there's
still
certainly
infrastructure
technology
that
you
know
we
answer
questions
about
on
a
regular
basis.
The
like,
if
you
wanted
this
right
now,
I,
could
have
something
going
very
quickly.
C
C
A
Well,
we're
an
open-source
project,
so
we
welcome
contributions
and
you
know
we're
we're
working
active
I,
mean
I,
think
the
open,
the
new
contributor
experience
is
made
much
easier
by
the
easier
installation.
I
think
right
now
we
have
a
very
we
have
a
significant
number
of
AP
is
that
could
that
are
described
and
linked
to
chaos?