New Data Connectors: JIRA and Google Analytics API connectors
have been added to the data connector list.
The ‘Notification’ feature has been provided for pushing
system notifications to users.
Audit Trail: Audit log/Performance Visualization has been
provided to the users.
Centralized Custom Attributes have been provided to manage the
bulk of users in one go.
Windows AD Attribute Synchronization.
The ‘Server Monitor’ module has been added to monitor server
operations.
Introduced new module to manage Data Sets and Data Services.
A new application ‘Play’ has been introduced to increase user
interactivity. Users can capture screen shots for multiple documents, annotate and
create an online slide presentation both on native iOS and the web.
A general workspace concept has been introduced for Dashboard
Designer, Play, and Data
Custom forms have been added to create and deploy custom
application inside the platform.
Enhancements:
Security Enhancements:
Encryption is added for secured data network
communication.
Service side permission checks can be done for data base
security.
Concurrency has been improved for rapid performance.
XSS HTML filter has been provided for the prevention of
malicious code injection.
UI enhancements for enriched look and feel of the Platform
modules like Administration, Data Center, and User Management.
Report plugin has been enhanced.
Note:
Data Store connector is currently in Beta.
The Play plugin is currently in Beta.
Business Story
New Features:
Introduced NLP driven text based data search feature.
Data Visualization Components
Data Grid - to view multiple data values in a tabular
format
Map - to position data in a geographical context using
different data layers
Enhancements:
Improved UI for enhanced user experience.
Data Preparation
New Features:
Extract: Equipped with ‘Full Load’ and ‘Incremental Load/Delta
Load’ functionality.
A range of components for transforming data:
Add Constants of all types to input datasets
Filter out rows from input Datasets by applying conditions
Perform Date Transformations on date columns
Perform ‘Search’ and ‘Replace’ functions on text columns.
Perform Computations involving Columns and Constants.
Select and Rename columns in the input dataset.
Convert one type of column to another type.
Perform SQL like Join operations on two input datasets.
Perform SQL like union operations on two input datasets.
Group by multiple columns and find Aggregates of columns.
Load Features- to write the transformed data set to elastic
search for visualization.
Workflow Features
Organize workflows into folders.
Schedule workflows to run at specified intervals.
Watch the status of the workflow executions.
Preview the results of the workflow executions before
scheduling runs.
Dashboard Designer
Charting New Features:
Charts:
Histogram
Waterfall
Decision Tree
Sparkline
The auto precision calculation has been added for charts.
Data point labels have been added for charts.
Map Chart: Addition of Australia, Canada, Asia, Europe, Middle
East, Asia Pacific maps
Text indicators are provided in the data grids.
Enhancements:
Bubble chart- Negative values and static radius option are
provided
The ‘Dual’ list filter has been added.
The ‘Multi Select’ list filter has been provided for the
smooth interaction with the ‘Checkbox’ component.
The Bar and Column chart components can display customized
stack size.
The Legends panel can be displayed as open by-default.
The default slider range size can be set as 100%/ 50%/ 25%/
auto for the Time Series chart.
An optional row has been added to display Sum/ Average / Count
of column values for the Data grids.
The Scorecard and pivot components can be associated with
checkbox-with-legend.
Multiple series are supported by the Funnel and Pyramid chart
components.
Strong colors (gradients) can be filled in the Heat Map chart.
The scale tick marks and precision have been added for the
Gauges.
Repeater chart’s title can accept fixed string and value from
the global variable.
The auto-manipulator grouping is available up to 2 levels in
the Group Bar chart.
Improved merged data operation performance by adding
UnderScore.js
Data grids now have an option to freeze the columns and set
the adjustable width to each column.
Export dashboard opens a model to control all or selected
components to export in pdf/ excel.
Designer New Features
Usability features
The properties panel can be opened by double click on
the component.
Component navigation is made easier by
inter-switchable Property palette/ Dataset palette and Script Editor.
Connection filter keys can be accessed from a
drop-down menu instead of writing manually.
inter-switchable configuration and script windows are
provided for a connection.
UI upgrades
Components can be grouped together and managed by the
object browser window.
Operations like ‘Rename Group’, ‘Show All / Hide All
Group Components’ are available.
Uniform look and feel for all model boxes in the
designer..
The ‘Help’ section has been added in the designer and
provided with more details on components/ configuration/ scripting methods
and sample dashboards.
The Dashboard has an option to preview dashboard
without opening in edit mode.
Enhancements:
A locked component in object browser can allow the users to
update all properties and assign data set except positioning and dimension
properties.
The dashboard loading has been improved by loading
configuration and property data files from the cache.
The ‘Calculated fields’ can be renamed to any valid name.
Timely refresh of connection is updated.
An option is added to the preference page, to hide all the
‘success/ error/ warning/ information’ notifications.
The ‘Summary’ can be displayed for the predictive workflows
inside the dashboard.
Predictive Workbench:
New Features:
Multiple workflows are now supported in R to simplify the
modeling process.
Components to support multiple workflows in R:
Data Preparation
Missing Value Replacement – Missing values in a
column can be replaced with a specified replacement options in data set.
Replacement Options: Mean, Median, Mode, Maximum, Minimum, remove entire
row, Remove entire column and Custom replacement)
R Split Data – Split the data set into train and
test based on a ratio with sampling method like Linear sampling,
Shuffled sampling, and Stratified sampling.
Algorithm
R CNR Tree – Validation provided to R-CNR Tree
to support train and test methodology.
R Naïve Bayes – Validation provided to R-Naïve
Bayes to support train and test methodology.
Validation
Cross Validation - Splits the data set based
on the selected number of folds.
Bootstrap – Splits the data set based on the
selected number of resampling.
Repeated cross validation – Repeats the split
based on the selected number of folds.
Leave one out cross validation – Splits data
based on leaving one data instance and constructs a model for all
other data instances to find the best model.
Apply Model
R Apply Model – User can connect upper part
with model and lower part with test/ actual data to get prediction
based on the model.
Save R Model – User can extract the best model
from executed workflow and save.
Performance
R performance – Users can compare up to three
workflows and evaluate the performance of the models to determine
the most suitable model. Users can set evaluation metrics as Binary
or Multi-class.
R Saved Models
List of saved models from where the user can
use it in the workflow.
The ‘Refresh’ button has been introduced to rerun the R
workflow with the latest data.
The ‘Clear Cache’ option is provided to clear the cached data.
Components introduced in Spark Environment(Alpha release):
Data Preparation
Spark Filter – Provides a space where users
can write condition to apply the dynamic filter.
Data Transformation
Spark PCA – Algorithm which can be used in
dimensionality reduction.
Spark Chi Square – to select features of
categorical data.
Spark Index to String – maps the column of
label indices to back to a column containing original labels.
Spark SQL Transformer – provides facility to
transform data based on SQL statement in spark.
Spark Group by – Aggregate result on basis of
column.
Algorithm
Classification
Spark Decision Tree – Decision Tree
with validation to introduce machine learning.
Spark Random Forest – Random Forest
with validation to introduce machine learning.
Recommendation Engine
Spark ALS – Introduced ALS to provide
recommendation Model.
Custom Scala Script
The ‘Custom Scala Script’ component has been
introduced to write a Scala script that can run in spark
Environment. Limitation – RDD with Mllib algorithms not supported.
Predictive services in Dashboard Designer – The predictive
workflows can be deployed on the dashboard designer to consume it as a data source
to display visualization.
Elastic Search Writer in Scheduler - Workflows can now be
scheduled and pushed to Business Story.
Auto numbering will be displayed for the alias names of all
the components of a workflow in the ‘Console’ tab while running a workflow.
The predicted column name will be auto-numbered.
The ‘Delta Load’ feature has been added in MYSQL to define a
primary key.
Enhancements:
Enhancements in Spark Components
Spark R Formula – Added formula text area.
Spark performance – The ‘Regression Evaluation Metrics’
has been added in Spark performance types.
Suitable alignment is provided for R summary option.
Icon enhancements have been done to distinguish between Spark,
R, and Java components.
Users can receive component specific messages to get informed
about the functionalities provided by the selected component.
The Scatter Matrix chart display is limited to 7*7 in K-Mean
algorithms.
Summary of all the components will be displayed through an
email while scheduling a workflow.
