Data scientists often face ambiguous challenges and, as a group, should use and make use of the design process to address these challenges. These slides briefly make the case for using the design process. Interested in more, reach out!
Strata preview 2014: Design thinking for dummies (data scientists)
1. design thinking for dummies (data scientists)
tuesday, february 11, 9:00 a.m.
@deanmalmgren
@mstringer
@laurieskelly
2014 february
strata preview
2. data scientists thrive with ambiguity
solve for x
project evolution
x=5+2
@deanmalmgren | bit.ly/design-data
3. data scientists thrive with ambiguity
solve for x
Ax=b
project evolution
x=5+2
@deanmalmgren | bit.ly/design-data
4. data scientists thrive with ambiguity
solve for x
Ax=b
project evolution
x=5+2
optimize
Ax=b
subject to
f(x) > 0
@deanmalmgren | bit.ly/design-data
5. data scientists thrive with ambiguity
solve for x
Ax=b
optimize
f(x)
project evolution
x=5+2
optimize
Ax=b
subject to
f(x) > 0
@deanmalmgren | bit.ly/design-data
6. data scientists thrive with ambiguity
solve for x
Ax=b
optimize
f(x)
optimize
“our profitability”
project evolution
x=5+2
optimize
Ax=b
subject to
f(x) > 0
@deanmalmgren | bit.ly/design-data
8. origins of ambiguity
unclear problems
identify the best locations to plant new trees
@deanmalmgren | bit.ly/design-data
9. origins of ambiguity
unclear problems
identify the best locations to plant new trees
how many?
what kinds of trees?
move old trees?
replace old trees?
@deanmalmgren | bit.ly/design-data
10. origins of ambiguity
unclear problems
identify the best locations to plant new trees
aesthetically pleasing?
maximize growth?
increase folliage?
offset CO2 emissions?
how many?
what kinds of trees?
move old trees?
replace old trees?
@deanmalmgren | bit.ly/design-data
11. “design process” is used everywhere
anticipate failure
generate
hypotheses
evaluate
feedback
1-4 week
iterations
build
prototype
@deanmalmgren | bit.ly/design-data
13. design and data science
challenges in practice
problem lost in translation
evaluate
feedback
generate
hypotheses
1-4 week
iterations
build
prototype
@deanmalmgren | bit.ly/design-data
14. design and data science
challenges in practice
problem lost in translation
generate
hypotheses
takes a long time to
collect data, analyze, and
build visualization
evaluate
feedback
1-4 week
iterations
build
prototype
@deanmalmgren | bit.ly/design-data
15. design and data science
challenges in practice
problem lost in translation
generate
hypotheses
takes a long time to
collect data, analyze, and
build visualization
evaluate
feedback
1-4 week
iterations
build
prototype
proof is in the pudding
@deanmalmgren | bit.ly/design-data
16. solve ambiguous problems
with an iterative approach
http://bit.ly/design-data
!
@deanmalmgren
dean.malmgren@datascopeanalytics.com