Joy Bonaguro explains her work as CDO at San Francisco City : role definition, open data & data science argumentation, being valuable for the people, starting with a problem... Almost as creating a startup inside a public institution that exists for about 200 years :)
7. We structure our goals to confront our challenges
Establish efficient
and effective data
governance
Make timely data
easily available
Increase use of data
in decision-making
12. Our SF Building Explorer integrates 12+ open datasets
to provide a holistic view of a building
13. The data silo challenge
– fragmented data governance and stewardship
– complex, time-consuming data integration or sharing
– complex regulatory control and data compliance
– data management and standards missing or inconsistent
23. What data?
Data Science and algorithmic
decision aides
Active performance
management
Endless, painful
reporting
Automated reporting
Benchmarks in data driven decision making
24.
25.
26. What data?
Data Science and algorithmic
decision aides
Active performance
management
Endless, painful
reporting
Automated reporting
Benchmarks in data driven decision making
27. Zachary
Kanin. The New
Yorker Collection/
The Cartoon Bank
Liberate employee capacity from the never-ending
reporting cycle
29. What data?
Data Science and algorithmic
decision aides
Active performance
management
Endless, painful
reporting
Automated reporting
Benchmarks in data driven decision making
33. We apply this framework to all of our programs (and
publish it)
34. What data?
Data Science and algorithmic
decision aides
Active performance
management
Endless, painful
reporting
Automated reporting
Benchmarks in data driven decision making
37. Introducing data science can evoke a range of reactions
Whatever, you’re
just being trendy.
Get out of our
business. We’re
doing this already.
Resource hog.
This will solve
everything!!!!
41. Service Issue:
Backlog is tackled via
first in, first out (FIFO)
What to
prioritize?
Data
Science
Service
Change
Data Science Process:
Create a model to
categorize and group
past and current cases
Service Change:
Prioritize cases based
on categories in order
of risk, need or
opportunity
Result: Department addresses high priority cases first
How it
works:
“Prioritize
your
backlog”
42.
43. Increased client list 200% for commercial and 1000% for multi-family
properties and created a new leading indicator to help prioritize those leads
44. Able to predict with 83%
accuracy who will drop out of by
13 months of child age
45.
46. Objective: Demonstrate power of data science
and eventually embed in culture and fabric
Develop Deliver Celebrate