06 Turbo Prep - Introduction Video Lecture | Rapid Miner: Learn and Master - Data & Analytics

[Music]
reminder turbo prep substantially
reduces the time spent on data
extraction transformation and loading it
also seamlessly integrates with rapid
miners Auto model as we'll show you in
this video if you're used to working in
rapid miner studio design view you can
start turbo prep by selecting it in the
views menu as you can see here all
options like transform cleanse generate
or model and process are grayed out and
inactive so the first thing to do is to
load some data we can do that by either
clicking on the text or the load data
button we can now choose to start with
one of our recently used data sets we
could import new data or we can access
data from other built-in options like
through the rapid miner repository which
is comparable to a local data folder
from the repository you also have the
option to retrieve data from a rapid
miner server repository or from a
database to which you may have
previously established a connection it
is noteworthy that if you select import
data you have the option to import data
from your from your local computer a
connected network drive or you could
retrieve it directly from a database
given that you have the relevant
connection details at hand for this
short introduction we want to load some
sample data so I'll choose the samples
folder here to open the data we could
double-click or simply select and then
click the load data button we're now
returning to the turbo prep start screen
and now all the available options are
active on the left hand side you'll see
five groups of actions summarizing all
kinds of data preparation steps you
could imagine you can start any of these
steps now then jump to another or repeat
a similar step within the same group you
can also do multiple steps while you're
in one of the data preparation sessions
let's do a simple example like renaming
a column to show this in more detail
renaming is a transformation my sample
data shows information on each passenger
on the titanic's famous maiden voyage to
transform a column you need to select it
and then choose the transformation i'll
change the name of the column sex to
gender and as i hit apply you can see
how the preview is updated we could now
continue to do more transformations but
we'll end this first preparation session
by clicking commit transformation now
let's try something a bit more
challenging like adding a column to do
that I'm choosing the generate group
because I'm generating a new column now
to calculate the size of the family of
each given passenger we could simply add
the number of siblings and the number of
parents plus 1 which is the passenger
her or himself all of this works nicely
with drag-and-drop and as I click update
preview I can now see this new column or
as we call it this new attribute appear
having done this fast and very intuitive
preparation step we can now also nicely
demonstrate the other available options
of this view by selecting history you
can now see what has been done and
committed so far and if needed you could
decide to rollback alternatively you
could branch off another preparation
process by copying the data after any
committed session turbo press not a
black box but instead you can select to
open the process which has been
generated for you in the background
while using the point-and-click
interface simply select process if you
want to review what has been done so far
you can go through each operator in the
process panel if you want to change the
calculation just select the generate
attributes operator and edit the
calculation but keep in mind that this
change is not going to propagate back
into turbo prep but only takes effect if
you run the process displayed here in
the design view now we can go back to
our turbo prep instead of continuing
with our menu items let's look briefly
at the table view you can see grey bars
which are histograms at the top of each
attribute indicating the distribution
characteristics of the data for the
gender there are two bars indicating
male and female for age you can see a
more detailed distribution underneath
each histogram there are colors
displayed which are indicators of the
quality to learn more on the column age
they'll right click which gives us a
handy context menu from here I could
decide to jump right into a session of
transformations or cleansing
alternatively I could sort my data
according to this column but keep in
mind that the sorting will only be
applicable to the current view not to
our actual data table for that you need
to select transformations then sort as
you can see this opens a transformation
session for us right there as a shortcut
you can also apply any change and see
how it effects your data if you don't
want to commit the change you can simply
discard them with cancel as we're doing
here now let's look at show details in
the column details for age you can see
the format which is numeric a big
display of the previously indicated
histogram some key statistics and of
course you can also find some
information on the quality as you can
see there are five quality indications
which rapid miner studio provides to you
to easily access the quality of any data
type the rows are categorized in red if
they are missing or infinity in light
blue 4i deenis in gray for stability and
the remaining values are marked green
you can find more details on the
different terms here
if you want to explore your data further
then you can select to open charts here
instead of saving this state of our data
set now to our repository we'll use it
right away to try and see if we can use
it to build a good model to predict the
survival of the Titanic passengers will
just use a temporary snapshot of the
data and then we walk through the
familiar auto model dialog steps I'll
uncheck the calculation intense ensemble
methods here as we just want to see the
mechanics
as we take a look at the easy-to-read
decision tree model we can see that our
added column is a useful criteria or
feature for the model to create a
decision path by the way what we just
did with creating this extra column is
also referred to as feature generation
if you combine this with selecting the
optimal set of input columns also known
as feature selection then you're even
doing feature engineering with this
hidden bonus lesson on data science I
will now end this tutorial thank you
very much for watching
[Music]
you