5. UnidosUS
UnidosUS, previously known as NCLR (National
Council of La Raza) is the nation’s largest
Hispanic civil rights and advocacy organization.
• Policy & Advocacy
• Programs
• Partner with affiliated local organizations that
work directly with community
• Affiliate Network – 280+ affiliates across the US
• Principal, Digital Services
• New role, started spring 2021
• Chief digital/data officer function
6. UnidosUS acts as Platform
We enable the
work of and act as
a convenor for
nonprofits working
with US Hispanic
community
TA + $
TA
+
$
Ecosystem of subgrants
and technical assistance
(TA) to affiliates and
other partners
7. Data is an overhead cost
DATA
DATA
Data is fragmented
and is a cost not an
asset.
Data approach is
tactical and
transactional.
Doesn’t support
ecosystem
8. Solution: UnidosUS JUNTOS PLATFORM
JUNTOS
PLATFORM
Affiliates /
Partners
Community
Government
“Virtual”
Affilliate
Affilliate 1
Affilliate 2
Affilliate 3
Funders
Data becomes an
organizational and
community
resource
Academia
Staff
DATA
DATA
10. High level data platform architecture
Data catalog
Data
Model
processing
Data Lake
(File level tagging)
monday
.com
Raisers
Edge
Other data
files: excel, csv,
etc
Concur
Read only
one way
exports
Intake
Processing
Data
Model
interface
JUNTOS
PLATFORM
Data Analyst Low
level queries using R,
SQL, Python, etc
Program Staff dashboards
WFD
Edu
Health
Finance
P&A
Housing
Data Warehouse
structured
“enterprise data model”
WFD
Edu
Housing
Finance
P&A
Health
MetaData
(Affiliate, Location, etc)
Analytics Data
Products
Adhoc &
scheduled
Queries
External Data
(Census, BLS, CDC, etc)
Reality
Alchemer
Quickbase
Great
plains
Housing
Health
Edu
WFD
Fiannce
P&A
12. Data people more important than data system
Right data person/Wrong system
Inefficient but effective
Wrong data person/Right system
Efficiently ineffective
Start with program person with “data mentality” who can be trained on data
tools/techniques rather than data person that needs to learn program.
13. Analytics does not replace Intuition
We hope to see something different from
a new data tool/approach because we
think the purpose of analytics is to replace
“subjective” human judgement
Purpose of analytics is to support human
work
• Data supporting communication
• Data supporting understanding
Intuition
Data
Analytics
Analytics Data
Intuition and data are part of a cycle, not two
different paths
Judgement
Judgement
Judgement
Intuition
14. Data Sophistication Journey
Stage Focus Question Tools Relevant and Easily Available
Immediate Fulfill
requests
Where is the data? understand the need to do more with data
but not exactly what that means
Email, phone Data doesn’t exist until we ask for it
Present Data
capture
What is the data? staff are able to ask simple questions about
the data because its easier to get to
Individual
databases
Data exists but not easily accessible
Past Reporting What is the data telling us about what is going on now?
standard metrics/business rules/reports are precalculated so
staff can spend time on more value adding data work
Data lake,
Data marts
Data exists and is easily accessible via
dashboards and other data tools
Future Analytics What is the data telling us about what we should do?
Data is a core part of how we plan out our work because it is
very easy to get to. Our educated guesses are now more data
supported, our gut feelings easily can be validated and fleshed
out with data analytics
Data
Warehouse,
Data Science
New data from analytics products is
easily created and easily accessible
Current
focus
An organization’s approach to data evolves over time
15. Digestibility – Meet your users where they are
Create appropriate level of data system for your data sophistication
Analytics cannot far outstrip your user’s ability to use it
Set expectations to right level – what is possible based on data quality etc
Don’t build a
Ferrari to go
through a field
16. “Data Entry Team” not “Users”
End users are fundamental to success but “Analytics" can be intimidating
Start your users slow
and work your way up
You
are
here
Start with the familiar
Recognizing
themselves in the
data increases
emotional buy-in
(eg GIS - first thing
people do a try and
find thier house)
17. Reduce Cognitive workload
Systems provide structure and orientation
• Logical flow (step 1, step 2, etc)
• Forest instead of trees (data points in
context)
18. Common language for communication
Goal is to have accurate metrics aligned with business goal
1. Cannot talk about accuracy if there isn’t agreement on the value being reported
2. Once the value is consistent, you can talk about if it’s accurate
3. Once it’s accurate you can talk about whether it’s relevant to business goal
Metric A
Report 1: 90
Report 2: 81
Report 3: 87
Metric A
Report 1: 87
Report 2: 87
Report 3: 87
Consistent
Metric A
Report 1: 85
Report 2: 85
Report 3: 85
Metric
aligned with
goal
Not
Relevant
Determine proposed change
and go through cycle again
Accurate Relevant
DATA KNOWLEDGE INFORMATION
Accuracy
Relevancy
Consistency
19. 0
1
2
3
4
5
6
Immediate Present Past Future
Data Confidence – Evolves over time
Confidence
starts off
unknown
Confidence drops but
uncertainty narrows as you
pull data into a reporting
system and you see your
actual data quality is poor
As data is actually
captured, our confidence
increases but is still
somewhat uncertain
Your understanding of
your data increases,
quality goes up, and
confidence rebounds
Set expectations when putting new system in place.
“Garbage in is Garbage out” – realization usually starts with seeing “Garbage out”
21. Not all interesting questions are relevant
Resources should be focused
on data that ultimately solves
the main problem of achieving
organizational goals
22. Prioritization
• Desired organizational success prioritize which outcomes business should focus on
• Desired business outcomes prioritize which decisions analytics should focus on
• Desired analytics decisions prioritizes which data reporting should focus on
Prioritize
Outcomes
Prioritize
Analytics
Prioritize
Data
23. Traceability
Traceability tells you what is impacted
by upstream mistakes or change in
business rules
System helps tie outputs back to
inputs
25. Data Science - Iteration
Science is usually
about incrementalism
not revolution
Maxwell Lorentz Michaelson, Morley
Poincaré
“Standing on the shoulders of giants”
26. Data Science - Experimentation
Independent Variables - controllable
Independent Variables - environmental Systems /
Processes
System helps to focus
on what should be
controlled and what of
that can actually be
controlled
27. Data Science - Statistics
Data as common language for
communication
• Stats vs individual anecdotes
• Shared institutional knowledge