1
00:00:07,620 --> 00:00:12,590
Since we’ve previously covered open source
tools for data science, let’s look at the

2
00:00:12,590 --> 00:00:17,110
commercial options you’ll find in many enterprise
projects.

3
00:00:17,110 --> 00:00:20,890
Take another look at the overview of different
tool categories.

4
00:00:20,890 --> 00:00:26,230
Since cloud products are a newer species,
they follow the trend of having multiple tasks

5
00:00:26,230 --> 00:00:28,710
integrated in tools.

6
00:00:28,710 --> 00:00:33,399
This especially holds true for the tasks marked
green in the diagram.

7
00:00:33,399 --> 00:00:38,450
Let’s start with the fully integrated visual
tools category.

8
00:00:38,450 --> 00:00:43,910
Since these tools introduce a component where
large scale execution of data science workflows

9
00:00:43,910 --> 00:00:50,379
happens in compute clusters, we’ve changed
the title here and added the word “Platform.”

10
00:00:50,379 --> 00:00:55,739
These clusters are composed of multiple server
machines, transparently for the user, in the

11
00:00:55,739 --> 00:00:57,159
background.

12
00:00:57,159 --> 00:01:02,780
Watson Studio, together with Watson OpenScale,
covers the complete development life cycle

13
00:01:02,780 --> 00:01:08,320
for all data science, machine learning, and
AI tasks.

14
00:01:08,320 --> 00:01:12,030
Another example is Microsoft Azure Machine
Learning.

15
00:01:12,030 --> 00:01:16,900
This is also a fully cloud-hosted offering
supporting the complete development life cycle

16
00:01:16,900 --> 00:01:22,030
of all data science, machine learning, and
AI tasks.

17
00:01:22,030 --> 00:01:28,130
And finally, another example is H2O Driverless
AI, which we’ve already introduced in the

18
00:01:28,130 --> 00:01:29,670
last video.

19
00:01:29,670 --> 00:01:34,920
Although it is a product that you download
and install, one-click deployment is available

20
00:01:34,920 --> 00:01:37,910
for the common cloud service providers.

21
00:01:37,910 --> 00:01:42,670
Since operations and maintenance are not done
by the cloud provider, as is the case with

22
00:01:42,670 --> 00:01:48,540
Watson Studio, Open Scale, and Azure Machine
Learning, this delivery model should not be

23
00:01:48,540 --> 00:01:56,780
confused with Platform or Software as a Service
-- PaaS or SaaS.

24
00:01:56,780 --> 00:02:03,400
In data management, with some exceptions,
there are SaaS versions of existing open source

25
00:02:03,400 --> 00:02:04,400
and commercial tools.

26
00:02:04,400 --> 00:02:09,410
Remember, SaaS stands for “software as a
service.”

27
00:02:09,410 --> 00:02:14,950
It means that the cloud provider operates
the tool for you in the cloud.

28
00:02:14,950 --> 00:02:20,900
As an example, the cloud provider operates
the product by backing up your data and configuration

29
00:02:20,900 --> 00:02:23,220
and installing updates.

30
00:02:23,220 --> 00:02:29,290
As mentioned, there is proprietary tooling,
which is only available as a cloud product.

31
00:02:29,290 --> 00:02:32,969
Sometimes it’s only available from a single
cloud provider.

32
00:02:32,969 --> 00:02:40,239
One example of such a service is Amazon Web
Services DynamoDB, a NoSQL database that allows

33
00:02:40,239 --> 00:02:46,590
storage and retrieval of data in a key-value
or a document store format.

34
00:02:46,590 --> 00:02:51,400
The most prominent document data structure
is JSON (pronounced “jay-sun”).

35
00:02:51,400 --> 00:02:56,540
Another flavour of such a service is Cloudant,
which is a database-as-a-service offering.

36
00:02:56,540 --> 00:03:02,010
But, under the hood it is based on the open
source Apache CouchDB.

37
00:03:02,010 --> 00:03:08,439
It has an advantage: although complex operational
tasks like updating, backup, restore, and

38
00:03:08,439 --> 00:03:14,079
scaling are done by the cloud provider, under
the hood this offering is compatible with

39
00:03:14,079 --> 00:03:15,209
CouchDB.

40
00:03:15,209 --> 00:03:21,260
Therefore, the application can be migrated
to another CouchDB server without changing

41
00:03:21,260 --> 00:03:22,510
the application.

42
00:03:22,510 --> 00:03:28,049
And IBM offers Db2 as a service as well.

43
00:03:28,049 --> 00:03:32,720
This is an example of a commercial database
made available as a software-as-a-service

44
00:03:32,720 --> 00:03:39,299
offering in the cloud, taking operational
tasks away from the user.

45
00:03:39,299 --> 00:03:44,200
When it comes to commercial data integration
tools, we talk not only about “extract,

46
00:03:44,200 --> 00:03:51,230
transform, and load,” or “ETL” tools,
but also about “extract, load, and transform,”

47
00:03:51,230 --> 00:03:53,980
or “ELT,” tools.

48
00:03:53,980 --> 00:03:59,669
This means the transformation steps are not
done by a data integration team but are pushed

49
00:03:59,669 --> 00:04:04,409
towards the domain of the data scientist or
data engineer.

50
00:04:04,409 --> 00:04:10,180
Two widely used commercial data integration
tools are Informatica Cloud Data Integration

51
00:04:10,180 --> 00:04:13,059
and IBM’s Data Refinery.

52
00:04:13,059 --> 00:04:19,079
Data Refinery enables transformation of large
amounts of raw data into consumable, quality

53
00:04:19,079 --> 00:04:23,130
information in a spreadsheet-like user interface.

54
00:04:23,130 --> 00:04:26,940
Data Refinery is part of IBM Watson Studio.

55
00:04:26,940 --> 00:04:32,930
The market for cloud data visualization tools
is huge, and every major cloud vendor has

56
00:04:32,930 --> 00:04:33,930
one.

57
00:04:33,930 --> 00:04:39,450
An example of a smaller company’s cloud-based
data visualization tool is DataMeer.

58
00:04:39,450 --> 00:04:46,449
IBM offers it’s famous Cognos Business intelligence
suite as cloud solution as well.

59
00:04:46,449 --> 00:04:52,729
IBM Data Refinery also offers data exploration
and visualization functionality in Watson

60
00:04:52,729 --> 00:04:53,759
Studio.

61
00:04:53,759 --> 00:05:00,330
Again, these are just some examples of a rapidly
changing and growing commercial ecosystem

62
00:05:00,330 --> 00:05:05,970
among a huge number of established and emerging
vendors.

63
00:05:05,970 --> 00:05:11,750
In Watson Studio, an abundance of different
visualizations can be used to better understand

64
00:05:11,750 --> 00:05:12,750
data.

65
00:05:12,750 --> 00:05:18,280
For example, this 3D bar chart enables you
to visualize a target value on the vertical

66
00:05:18,280 --> 00:05:24,060
dimension, which is dependent on two other
values on the horizontal dimensions.

67
00:05:24,060 --> 00:05:28,930
Coloring enables you to visualize a third
dimension.

68
00:05:28,930 --> 00:05:35,470
Hierarchical edge bundling enables you to
visualize correlations and affiliations between

69
00:05:35,470 --> 00:05:36,470
entities.

70
00:05:36,470 --> 00:05:42,409
If sufficient, a classic bar chart can do
the job as well, whereas a 2D scatter plot

71
00:05:42,409 --> 00:05:50,729
with a heat map shows two dependent data fields,
one on the y axis and one as color intensity.

72
00:05:50,729 --> 00:05:57,080
A tree map shows distribution of subsets within
a set, the famous pie chart does the same

73
00:05:57,080 --> 00:06:03,610
but in a non-hierarchical manner, and finally,
a word cloud pops out significant terms in

74
00:06:03,610 --> 00:06:05,870
a document corpus.

75
00:06:05,870 --> 00:06:10,310
Model building can be done using a service
such as Watson Machine Learning.

76
00:06:10,310 --> 00:06:16,379
Watson Machine Learning can train and build
models using various open source libraries.

77
00:06:16,379 --> 00:06:21,500
Google has a similar service on their cloud
called AI Platform Training.

78
00:06:21,500 --> 00:06:26,949
Nearly every cloud provider has a solution
for this task.

79
00:06:26,949 --> 00:06:32,199
Model deployment in commercial software is
usually tightly integrated to the model building

80
00:06:32,199 --> 00:06:33,250
process.

81
00:06:33,250 --> 00:06:38,879
Here is an example of the SPSS Collaboration
and Deployment Services, which can be used

82
00:06:38,879 --> 00:06:44,639
to deploy any type of asset created by the
SPSS software tools suite.

83
00:06:44,639 --> 00:06:47,240
The same holds for other vendors.

84
00:06:47,240 --> 00:06:52,280
In addition, commercial software can export
models in an open format.

85
00:06:52,280 --> 00:06:58,670
As an example, SPSS Modeler supports exporting
models as Predictive Model Markup Language,

86
00:06:58,670 --> 00:07:05,970
or “PMML,” which can be read by numerous
other commercial and open software packages.

87
00:07:05,970 --> 00:07:11,340
Watson Machine Learning can also be used to
deploy a model and make it available to consumers

88
00:07:11,340 --> 00:07:13,950
using a REST interface.

89
00:07:13,950 --> 00:07:19,479
Amazon SageMaker Model Monitor is an example
of a cloud tool that continuously monitors

90
00:07:19,479 --> 00:07:23,340
deployed machine learning and deep learning
models.

91
00:07:23,340 --> 00:07:27,820
Again, every major cloud provider has similar
tooling.

92
00:07:27,820 --> 00:07:31,310
This is also the case for Watson OpenScale.

93
00:07:31,310 --> 00:07:33,289
OpenScale and Watson Studio…

94
00:07:33,289 --> 00:07:35,520
…unify the landscape.

95
00:07:35,520 --> 00:07:40,139
Everything marked in green can be done using
Watson Studio and Watson OpenScale.

96
00:07:40,139 --> 00:07:44,020
We’ll cover Open Scale will be covered in
a later video.

97
00:07:44,020 --> 00:07:49,039
You’ve learned how the most common tasks
in data science are supported by commercial

98
00:07:49,039 --> 00:07:51,340
cloud tools.

99
00:07:51,340 --> 00:07:56,650
Integration provides us the ability to use
the same tools for multiple tasks.

100
00:07:56,650 --> 00:08:03,319
In the next videos, we’ll look at packages,
APIs, datasets, and models for data science.