LEX 44909 second tranche Redacted.pdf

114
Talking points: meeting, 10 May 2021

The wider context:
  Data-sharing and data-driven policy is on the national agenda.
o  This work can position your portfolio at the cutting edge of using data to drive effective programs and
policy.
  Economic recovery is a core priority.
o  This work will contribute to the body of evidence on the impacts of 2020 economic shocks.
o  We will be building data assets and IT infrastructure to generate evidence on an ongoing basis on
social harm, welfare dependence and disadvantage.

Main objectives of the CDC data catalogue and data analytics (May 2021–June 2022):
  Key questions:
o  What do we know about social harms in CDC regions?
o  When is CDC an effective and appropriate way to deliver welfare?
  Generate evidence to contribute to the CDC future state
  Long-term objective: build data assets and capabilities to inform policy into the future

Proposed timeline for building the data catalogue and generating findings:
  31 Dec. 2021: preliminary findings (descriptive analysis)
  31 Mar. 2022: first phase of findings available
  30 June 2022: second phase of findings available

How we will reach these milestones:
  Access data assets to build the data library: Commonwealth, state/territory, and private data
o  Commenced negotiations to access data assets: this will be an ongoing project
o  New data assets added to data library as access secured
o  Data assets will be linked where possible and data regularly refreshed.
  Expand IT infrastructure and capabilities (including SAS)
  Data analytics:
o  Commencing analysis of linked welfare payment and CDC data (internal departmental analysis)
o  Option to commission comparative analysis of spending data, comparing CDC and other income support
recipients (would be commissioned from AlphaBeta)
o  Multiple strands of analysis using linked state/territory and Commonwealth data assets (using consultant
as recommended in MS21-000173)
  All analysis will be re-run at regular intervals, continuing to generate new findings.
  Findings and evidence:
o  Visualise findings: create digital dashboards by default
o  Share findings: demonstrate benefits to stakeholders
o  Evidence will inform CDC future state

Accessing state/territory and private data assets will be challenging:
  Dependent on negotiating access to data; need to manage governance, privacy and storage requirements
  Our strategy:
o  Build a reputation for best-practice management and use of data
o  Demonstrate our capability: how we can use data to understand and target social harms
o  Leverage this to negotiate further data access

115
Questions and answers

Q:  Why is the department recommending engaging a consultant?
A:  We need to build a data library and data infrastructure. We aren’t procuring a report or an evaluation; we are
procuring services to build out capability to analyse data on CDC on an ongoing basis.

With a data library, IT infrastructure and analytics models, the department will be able to conduct analysis on a
rolling basis. This means that we will be able to use the latest data to generate new evidence on social harms and
the CDC as a way to deliver welfare payments, and we will be able to do so whenever it is needed.

Q:  Why not just use AIHW data? Why not just commission AIHW to undertake analysis?
A:  First, the AIHW data assets were not built to answer CDC policy questions — so they are unlikely to have the most
directly relevant data. Second, commissioning analysis from AIHW would likely require more time to generate
findings. Third, AIHW’s services would also be one-off: we would receive a static report.
We recommend commissioning a consultant to build the department’s capabilities to generate evidence of
impact on an ongoing basis. Building our own data library will ensure that we have ongoing access tp the data
assets that are most relevant to CDC outcomes. Commissioning data analytics will allow us to re-run analysis as
needed, rather than having a single report that cannot be updated.

Q:  Why is new data infrastructure needed?
A:  Data infrastructure is needed to support the data library and allow us to run analysis on a recurring basis.

Q:  Does the department already have the data that will be used? If so, what are you doing with it now? If not,
how long will it take to get access?
A:  The department has some data assets, but we are aiming to build a much more comprehensive data library.
We have linked CDC data with welfare payment data and are commencing analysis; at the same time, we are
concurrently pursuing additional data assets.
Based on current discussions, we anticipate that we will be able to access some data assets fairly quickly. Other
negotiations will be more complex. Our strategy is to position the department as a best practice user of data and
to demonstrate what we are able to do with shared data. Over time, we expect that this will allow us to secure
agreement from a wider range of institutions.

Q:  Why does it take time to get access to data assets?
A:  There are legal and regulatory requirements governing access, storage, and use of data assets. These
requirements vary depending on the nature of the data (such as whether it is personal or sensitive), and vary
across jurisdictions. It isn’t possible to circumvent these processes.

Q:  What will you do if you can’t get access to data assets?
A:  There are a large number of data assets that we are seeking to negotiate access to: the ‘snowflake’ diagrams
show the variety of institutions holding data that we may be able to use. We are currently discussing data access
with a selection of these. If initial discussions are not successful, we will commence discussions with other
institutions.

Q:  Will the department be able to commission analysis of spending data (both CDC spending data and external
data assets such as the AlphaBeta spending data)?
A:  The department has commenced scoping this, and anticipates that this will be possible. The timeframe will
depend on the applicable data governance processes: we will provide an update on this by Wednesday.

117
Objectives
Better use of data:
• Access administrative data, including data held by states and territories
• Expand capability to generate evidence on an ongoing basis
Generate evidence
• Introduction of legislation anticipated in Spring 2022
• Future state of CDC
Proof of concept: use of integrated data to inform policy

121
Possible data sources: agencies and types of data

122
Possible data sources: agencies and types of data

123
Possible data sources: agencies and types of data

124
Possible data sources: agencies and types of data

125
Implementation
• Data infrastructure: to hold data assets that can be used to create baselines and
build an evidence base
• Data acquisition: data-sharing agreements with states and territories; accessing
data held by other Commonwealth agencies
• Analytics: data catalogue which wil serve as an inventory, to house al CDC data
assets
31 Oct. 2021
31 Dec. 2021
31 Mar. 2021
30 Jun. 2022
Demonstrate approach
Preliminary
Second phase
Third phase
and findings to be
findings and
findings
findings
Continue
generated (‘mock-up’):
progress
available
available
expanding
inform approach in
update
capabilities
engaging col eagues

126
Data infrastructure
The Cashless Welfare Engagement and Support Services Branch is
developing a data asset to enhance its data analytic capabilities.
• The department has secured a team of technical specialists from Deloitte to
support the project work and design infrastructure holdings for the Cashless
Debit Card and Income Management programs and build a baseline data
asset.
• Four members of the specialist team have commenced with the department just over
five weeks ago.
• The specialists are deep into the discovery phase of the project. The next phase includes
building and delivering.
• Al phases are currently on track

127
Talking points: meeting, 18 August 2021
Main objectives of the CDC data catalogue and data analytics (May 2021–June 2022):
 Key questions:
o
What do we know about social harms in CDC regions?
o
When is CDC an effective and appropriate way to deliver welfare?
 Generate evidence to contribute to the CDC future state
o
Noting legislation sunsets in Dec. 2022
 Long-term objective: build data assets and capabilities to inform policy into the future
 Proof of concept: use of integrated data to inform policy
Proposed timeline for building the data catalogue and generating findings:
 31 Oct. 2021: demonstrate approaches and types of findings to be generated (previously referred to mock-up)
and provide some early data to help inform policy direction and engagement with Cabinet colleagues
 Dec 2021: preliminary findings and update
 31 Mar. 2022: second phase of findings available
 30 June 2022: third phase of findings available
 These are ambitious timelines: the scope of the work is considerable.
o
The work is heavily dependent on data integration capability and capacity.
o
The department is carefully managing this project and will alert the office if there are risks to delivery.
How we will reach these milestones:
 Access data assets to build the data library: Commonwealth, state/territory, and data held by private
organisations (NGOs, research orgs etc.)
o
Commenced negotiations to access data assets: this will be an ongoing project
o
New data assets added to data library as access secured with an initial pilot with one to two states
(WA/NT)
o
Data assets will be linked where possible and data regularly refreshed.
 Expand data infrastructure and capabilities
 Data analytics:
o
Commencing analysis of linked welfare payment and CDC data (internal departmental analysis)
o
Multiple strands of analysis using linked state/territory and Commonwealth data assets (using consultant
as recommended in MS21-000173)
 All analysis will be re-run at regular intervals, continuing to generate new findings.
 Findings and evidence:
o
Visualise findings: create digital dashboards by default
o
Share findings: demonstrate benefits to stakeholders
o
Evidence will inform CDC future state
Accessing state/territory and private data assets — managing risks to delivery:
 Dependent on negotiating access to data; need to manage governance, privacy and storage requirements
 Our strategy:
o
Build a reputation for best-practice management and use of data
o
Demonstrate our capability: how we can use data to understand and target social harms
o
Leverage this to negotiate further data access

128
Data linkage, privacy and security:
 All integration of state/territory data will comply with all relevant regulations and legislation.
 Data will be de-identified and assigned unique match key. This will ensure any privacy or confidentiality concerns
are managed.
 Data will be handled and stored in secure IT environments with access controls managed by the relevant data
custodians.
s47B
Questions and answers
Q:  Why was new data infrastructure needed?
A:  Data infrastructure is needed to support the data library and allow us to run analysis on a recurring basis.
Q:  Why did the department engage a technical specialists?
A:  We needed to build a data library and data infrastructure. We weren’t procuring a report or an evaluation; we
are procured services to build our capability to analyse data on CDC on an ongoing basis.
With a data library, IT infrastructure and analytics models, the department will be able to conduct analysis on a
rolling basis. This means that we will be able to use the latest data to generate new evidence on social harms and
the CDC as a way to deliver welfare payments, and we will be able to do so whenever it is needed.
Q:  What steps are being taken to access administrative data?
A:  We are focussed on improving the use of quantitative administrative data, with priority given to data held by
states and territories.
Given the legislation sunset, we are focussed on ensuring that evidence to be available for stakeholder
engagement prior to the introduction of legislation next year.
We imagine that you may also want data to inform decisions about the future state of the CDC – for instance if
you wish to consider refining policy settings such as targeting.

129
Q:  Does the department already have the data that will be used? If so, what are you doing with it now? If not,
how long will it take to get access?
A:  The department has some data assets, but we are aiming to build a much more comprehensive data library.
We have linked CDC data with welfare payment data and are commencing analysis; at the same time, we are
concurrently pursuing additional data assets.
Q:  Why does it take time to get access to data assets?
A:  There are legal and regulatory requirements governing access, storage, and use of data assets. These
requirements vary depending on the nature of the data (such as whether it is personal or sensitive), and vary
across jurisdictions. It isn’t possible to circumvent these processes.
Q:  What will you do if you can’t get access to data assets?
A:  There are a large number of data assets that we are seeking to negotiate access to: the ‘snowflake’ diagrams
show the variety of institutions holding data that we may be able to use. We are currently discussing data access
with a selection of these. If initial discussions are not successful, we will commence discussions with other
institutions.
Q: Is it possible to leverage new processes linked to the Intergovernmental Agreement of data sharing and the
national data priority areas?
A:  It’s correct to say that data-sharing related to the Intergovernmental Agreement is a priority for states and
territories, along with the Data and Digital Ministers’ Meeting.
The national data priority areas, however, are major policy priorities with multiple projects to be pursued under
each. The CDC work is currently more focussed. We recommend that you engage with your Ministerial colleagues
outside the Intergovernmental Agreement processes.
These processes linked to the Intergovernmental Agreement also have a high profile, and fewer opportunities to
understand and manage stakeholder concerns. The department is being strategic in engaging with stakeholders
in CDC regions in order to ensure that we can demonstrate that we are following best practice in how we use
data and address any concerns from CDC participants.
Q:  Why not just use AIHW data? Why not just commission AIHW to undertake analysis?
A:  First, the AIHW data assets were not built to answer CDC policy questions — so they are unlikely to have the most
directly relevant data. Second, commissioning analysis from AIHW would likely require more time to generate
findings. Third, AIHW’s services would also be one-off: we would receive a static report.
We have commissioned a technical specialist to build the department’s capabilities to generate evidence of
impact on an ongoing basis. Building our own data library will ensure that we have ongoing access to the data
assets that are most relevant to CDC outcomes. Commissioning data analytics will allow us to re-run analysis as
needed, rather than having a single report that cannot be updated.

130
Cashless Debit Card Data Analytics
Capability Uplift Plan
Department of Social Services
Cashless Welfare Engagement and Support Services

131
Contents
1.
Purpose
2
2.
Learning Objectives & Requirements
2
2.1
Required Knowledge
2
2.1.1
SAS Required Knowledge
2
2.1.2
SQL Required Knowledge
3
2.1.3
Non-technical Skills
3
2.2
Current Skill & Knowledge Level
3
3.
Training Plan
5
3.1
Approach
5
3.2
Mandatory Training
5
3.2.1
SAS Mandatory Training
5
3.2.2
SQL Mandatory Training
6
3.3
Data Analytics Workshops
6
3.4
Ways of Working
7

132
1. Purpose
In support of The Department of Social Services (the Department) continued investment in Cashless
Welfare, Deloitte has created a Data Asset (the Asset) for the Cashless Welfare Engagement and
Support Services (CWESS) Branch. The Asset has provided the branch the ability to access and curate
historical data related to the Cashless Debit Card (CDC) Program.
To ensure continued use and value derived from this project, Deloitte has developed the following
capability uplift plan to ensure the Department has the capability to analyse and derive insights into
the CDC Program. Analysis of the Asset and state or federal government data sources, especially
integrated, requires a level of knowledge and skill to avoid misrepresentations and misinterpretations
of data.
2. Learning Objectives & Requirements
2.1 Required Knowledge
To derive insights and an understanding of how the CDC is impacting relevant communities, analysts
within the CWESS branch must have an understanding of the data, analytics tools and systems
available within the Department. Analyst must have a technical understanding of the Assets software
packages (SAS) and programming languages (SQL), as well as a foundational knowledge of data
analytics techniques.
2.1.1 SAS Required Knowledge
SAS is the core software package utilised within the Data Asset. It gives analysts the ability to curate,
develop and analyse data sets available to the branch to derive insights into the impact of policy in
Australian communities. Additionally, the tools available within SAS Enterprise Guide (EG) gives the
branch the ability to visualise curated data to identify immediate trends and additional avenues of
analysis.
To utilise this capability available, analysts must have an understanding of the following SAS functions
and code:
 Data [step]
 FIRST; LAST
 PROC DATASETS
 PROC SORT
 PROC SQL
 FROMAT
 IF; THEN
 %LET
 PROC REPORT
 RETAIN
 %SETDATE
Analyst should also be familiar with the following SAS EG summary and visualisation options available
to them. Additionally, users must also become adept at modifying their own programs to change
additional parameters and visual features such as colours.
 SUMMMARY STATISTICS
 LINE PLOT
 SUMMARY TABLES
 SCATTER PLOT
 BAR CHART
 BAR-LINE CHART

133
2.1.2 SQL Required Knowledge
SQL (Structured Query Language) Is the industry standard programming language for interfacing with
Relational Databases. This is core knowledge for analysts as it allows for insert, search, update and
delete database records. For SQL, we specifically require the Data Analytics team to be able to
perform the following functions:
 Queries and sub-queries
 CASE WHEN
 INTCK
 FORMAT
 CREATE TABLE; CREATE
 GROUP BY
 INTNX
VIEW
 UPCASE
 ORDER BY
 ROUND
 Filter in query:
 Calculations;
 JOIN
o
WHERE
o
MAX/MIN
o
Left
o
DISTINCT
o
FREQ/RANGE
o
Right
o
HAVING
o
SUM/COUNT
o
Cross
o
=,<>,<,>,>=,=<,IN
o
AVG
o
Full
[LIST], BETWEEN, IS
o
ABS
o
Inner
MISSING, IN A RANGE,
IS NULL, CONTAINS,
LIKE
2.1.3 Non-technical Skills
As well as having the technical skills required to undertake advanced analytics within the Asset,
analysts must also have an understanding of foundational statistics. Statistical literacy is vital to being
able to analyse and interpret results accurately. Analysts should understand which method they are
employing, such as sum, mean/average, standard deviation, range, min, max, quartiles, frequency,
number of missing values, weighted average.
In addition to statistical literacy, analysts require an understanding of the data within the Asset and
their foundation of rules and logic beginning with Services Australia. Since the first trial site began in
2016, there have been numerous changes to policy, benefits, and data structuring; while complete
knowledge of Services Australia and its specific data warehouse is not necessary, users must be able
to identify abnormalities in the data, which may indicate that the Asset requires an update or
maintenance. Understanding the rules behind column values prevents analysts from misinterpreting
and misrepresenting the data. For example, it is important that an analyst knows when a participant is
‘off’ the program, or just paused.
2.2 Current Skill & Knowledge Level
Through working alongside the Data Analytics team to produce the October Ministers Office (MO)
update and December MO update Deloitte has seen a range of skill levels and comprehension within
the Data Analytics team. We have made the following assessment of the current ability of the team as
a whole in relation to the core knowledge required to produce analytic outputs for reporting.