clarifying a few sections in labs

This commit is contained in:
furmangg 2021-10-12 17:38:52 -05:00
Родитель 8658f3187b
Коммит c7d49afe60
5 изменённых файлов: 17 добавлений и 14 удалений

Двоичные данные
images/lab04/media/imageCleanMissingDataSearch.png Normal file

Двоичный файл не отображается.

После

Ширина:  |  Высота:  |  Размер: 12 KiB

Просмотреть файл

@ -102,7 +102,7 @@ and wish take advantage of a trial environment, visit
<img src="images/lab01/media/image3.png" style="width:7.5in;height:4.46597in" alt="Graphical user interface, application, Teams Description automatically generated" />
1. Choose “Import data” and enter a dataset name “CustomerDemographics” and click Next.
1. Choose “Import data” (_Note: the UI has been renamed to "Microsoft Power Query" in the latest release_) and enter a dataset name “CustomerDemographics” and click Next.
<img src="images/lab01/media/image4.png" style="width:7.5in;height:4.73333in" alt="Graphical user interface, text, application, email Description automatically generated" />
@ -137,7 +137,7 @@ and wish take advantage of a trial environment, visit
<img src="images/lab01/media/image10.png" style="width:7.5in;height:3.9875in" alt="Graphical user interface, application, table Description automatically generated" />
1. Choose to refresh manually (since this is a static sample dataset that
1. Choose "refresh manually" (since this is a static sample dataset that
wont change) and click Save.
<img src="images/lab01/media/image11.png" style="width:7.5in;height:4.53125in" alt="Graphical user interface, text, application Description automatically generated" />
@ -149,7 +149,7 @@ and wish take advantage of a trial environment, visit
## Step 3: Load CustomerServiceAttributes sample dataset to Customer Insights
1. Click “Add data source”, Import data, and name it
1. Click “Add data source”, Import data (_Note: the UI has been renamed to "Microsoft Power Query" in the latest release_), and name it
CustomerServiceAttributes and click Next:
<img src="images/lab01/media/image13.png" style="width:7.5in;height:4.67986in" alt="Graphical user interface, text, application, email Description automatically generated" />

Просмотреть файл

@ -197,10 +197,7 @@ designing the experiment.
6. In the Advanced Settings, you will see your selections from the
prior screen at the top. Name the compute **CustomerChurnCC**. Set
the Minimum number of nodes at 0 and Maximum at 2. Set the Idle
seconds before scale down to 60. Click Create when complete.
**Important: This will be mentioned later on as well, but we need to
make sure and scale down our clusters when were done to avoid
incurring charges.**
seconds before scale down to 60. Click Create when complete and wait a few minutes.
<img src="images/lab04/media/image22.png" style="width:6.14467in;height:2.1375in" alt="Graphical user interface, text, application Description automatically generated" />
@ -239,12 +236,16 @@ designing the experiment.
11. Next, we know that TotalCharges has a few blank values (we could
have tested this via using the Summarize Data asset and evaluating
each column distribution but for the sake of this lab were going to
move on to handling them). Drag **Clean Missing Data** onto the page
move on to handling them). Search for **Clean Missing Data**:
<img src="images/lab04/media/imageCleanMissingDataSearch.png"/>
Drag **Clean Missing Data** onto the page
and connect TelcoChurn to the new asset.
<img src="images/lab04/media/image27.png" style="width:2.70438in;height:1.87083in" alt="A picture containing chart Description automatically generated" />
12. In the properties pane for Clean Missing Data. Select TotalCharges
12. In the properties pane for Clean Missing Data. Click "Edit column" and select TotalCharges
as the Column and set the parameters below. *Note: If the goal of
the lab was to build the best model we would have run some form of
regression or sub-sample from the complete rows to find a predicted
@ -290,7 +291,7 @@ designing the experiment.
15. The right node of the Split Data asset is connected to another
instance of **Select Columns in Dataset**. This extracts the
Customer IDs of the testing data. We are going to join this back in
to the results dataset after the models have run.
to the results dataset after the models have run. _Note that the description "Extract customer ID to join back in later" can be added in the Comment section of the properties pane of that module. This is optional._
<img src="images/lab04/media/image33.png" style="width:3.04853in;height:1.57265in" alt="Diagram Description automatically generated" /><img src="images/lab04/media/image34.png" style="width:3.46568in;height:1.52663in" />
@ -315,6 +316,8 @@ designing the experiment.
## Step 4: Training models
_Note: If you get confused on how modules should be connected, feel free to skip down to a zoomed-out screenshot in Step 5.7 below._
Now that weve setup the design of our experiment, we can begin to train
the models and evaluate the outputs.

Просмотреть файл

@ -80,7 +80,7 @@ learning models into Customer Insights.
<img src="images/lab05/media/image6.png" style="width:7.5in;height:3.025in" alt="Graphical user interface, text, application, email Description automatically generated" />
1. Choose Import data and give it the name “AzureMlResults” and click Next.
1. Choose Import data (_Note: the UI has been renamed to "Microsoft Power Query" in the latest release_) and give it the name “AzureMlResults” and click Next.
<img src="images/lab05/media/image7.png" style="width:7.5in;height:5.04792in" alt="Graphical user interface, text, application, email Description automatically generated" />

Просмотреть файл

@ -209,12 +209,12 @@ flow is executed under the covers on highly scalable Spark compute.
<img src="images/lab06/media/image24.png" style="width:1.86484in;height:0.30213in" />
1. Click the Import schema button now. Accepting the default formats is
adequate for our file:
1. Click the Import schema button (which was grayed out in step 13) now. Accepting the default formats is
adequate for our file. Click the Import button at the bottom to accept the defaults:
<img src="images/lab06/media/image25.png" style="width:6.40714in;height:5.01112in" alt="Table Description automatically generated" />
1. Now the schema for this CustomerChargesDetail source is populated.
1. Wait a few seconds for the schema to be loaded. Now the schema for this CustomerChargesDetail source is populated.
<img src="images/lab06/media/image26.png" style="width:7.5in;height:2.86389in" alt="Graphical user interface, application Description automatically generated" />