Data Science at Berkeley
7 likes6,605 views
This document provides an overview of data science education at UC Berkeley presented by Joshua Bloom. It discusses the core principles of data science as a discipline and debates whether it should be taught as an academic subject or skillset. It also describes Berkeley's efforts to teach data science through bootcamps, seminars, and degree programs. Examples are given of students' projects applying machine learning and Bayesian techniques to astronomy and other domains. The document advocates teaching data literacy before big data proficiency and stresses the importance of a well-rounded toolbox and hands-on training for modern data-driven scientists.
![mij = µi + M0j
+ ↵j log10 (Pi/P0)
+ E(B V )i ⇥ [RV ⇥ a (1/ i) + b (1/ i)]
+ ✏ij
Bayesian$Distance$Ladder
Some-Variable-Stars-show-a-PeriodDBrightness-CorrelaFon
i'indexes'over'individual'stars
j'indexes'over'wavebands
a'and'b'are'fixed'constants'at'each'color
Data--134'RR'Lyrae'(ultraviolet'to'infrared)
Fit--307-dimensional-model-parameter-inference
K'determinisHc'MCMC'model'with'PyMC
K'~6'days'for'a'single'run'(one'core)
K'parallelism'for'convergence'tests
Klein+12;'Klein,JSB+14,
“Leavitt Law”
HenrieKa-Swan-LeaviK](https://image.slidesharecdn.com/pydatabloom16x9-140504231706-phpapp02/85/Data-Science-at-Berkeley-7-320.jpg)
![Data-Centric Coursework, Bootcamps, Seminars, & Lecture
Series
BDAS: Berkeley Data Analytics
Stack
[Spark, Shark, ...]
parallel
programming
bootcamp
...and entire degree programs](https://image.slidesharecdn.com/pydatabloom16x9-140504231706-phpapp02/85/Data-Science-at-Berkeley-16-320.jpg)
![Python for Data Science @ Berkeley [Sept 2013]](https://image.slidesharecdn.com/pydatabloom16x9-140504231706-phpapp02/85/Data-Science-at-Berkeley-24-320.jpg)
![Time domain preprocessing
- Start with raw photometry!
- Gaussian process detrending!
- Calibration!
- Petigura & Marcy 2012!
!
Transit search
- Matched filter!
- Similar to BLS algorithm (Kovcas+ 2002)!
- Leverages Fast-Folding Algorithm
O(N^2) → O(N log N) (Staelin+ 1968)!
!
Data validation
- Significant peaks in periodogram, but
inconsistent with exoplanet transit
TERRA – optimized for small planets
Detrended/calibrated photometry
TERRA
RawFlux(ppt)CalibratedFlux
Erik Petigura
Berkeley Astro
Grad Student
Prevalence of Earth-size planets orbiting Sun-like stars
Erik A. Petiguraa,b,1
, Andrew W. Howardb
, and Geoffrey W. Marcya
a
Astronomy Department, University of California, Berkeley, CA 94720; and b
Institute for Astronomy, University of Hawaii at Manoa, Honolulu, HI 96822
Contributed by Geoffrey W. Marcy, October 22, 2013 (sent for review October 18, 2013)
Determining whether Earth-like planets are common or rare looms
as a touchstone in the question of life in the universe. We searched
for Earth-size planets that cross in front of their host stars by
examining the brightness measurements of 42,000 stars from
National Aeronautics and Space Administration’s Kepler mission.
We found 603 planets, including 10 that are Earth size (1 − 2 R⊕)
and receive comparable levels of stellar energy to that of Earth
(0:25 − 4 F⊕). We account for Kepler’s imperfect detectability of
such planets by injecting synthetic planet–caused dimmings into
the Kepler brightness measurements and recording the fraction
detected. We find that 11 ± 4% of Sun-like stars harbor an Earth-
size planet receiving between one and four times the stellar inten-
sity as Earth. We also find that the occurrence of Earth-size planets is
constant with increasing orbital period (P), within equal intervals of
logP up to ∼200 d. Extrapolating, one finds 5:7+1:7
−2:2 % of Sun-like stars
harbor an Earth-size planet with orbital periods of 200–400 d.
extrasolar planets | astrobiology
The National Aeronautics and Space Administration’s (NASA’s)
Kepler mission was launched in 2009 to search for planets
that transit (cross in front of) their host stars (1–4). The resulting
dimming of the host stars is detectable by measuring their bright-
ness, and Kepler monitored the brightness of 150,000 stars every
30 min for 4 y. To date, this exoplanet survey has detected more
than 3,000 planet candidates (4).
We searched for transiting planets in Kepler brightness mea-
surements using our custom-built TERRA software package
described in previous works (6, 9) and in SI Appendix. In brief,
TERRA conditions Kepler photometry in the time domain, re-
moving outliers, long timescale variability (>10 d), and systematic
errors common to a large number of stars. TERRA then searches
for transit signals by evaluating the signal-to-noise ratio (SNR) of
prospective transits over a finely spaced 3D grid of orbital period,
P, time of transit, t0, and transit duration, ΔT. This grid-based
search extends over the orbital period range of 0.5–400 d.
TERRA produced a list of “threshold crossing events” (TCEs)
that meet the key criterion of a photometric dimming SNR ratio
SNR > 12. Unfortunately, an unwieldy 16,227 TCEs met this cri-
terion, many of which are inconsistent with the periodic dimming
profile from a true transiting planet. Further vetting was performed
by automatically assessing which light curves were consistent with
theoretical models of transiting planets (10). We also visually
inspected each TCE light curve, retaining only those exhibiting a
consistent, periodic, box-shaped dimming, and rejecting those
caused by single epoch outliers, correlated noise, and other data
anomalies. The vetting process was applied homogeneously to all
TCEs and is described in further detail in SI Appendix.
To assess our vetting accuracy, we evaluated the 235 Kepler
objects of interest (KOIs) among Best42k stars having P > 50 d,
which had been found by the Kepler Project and identified as planet
candidates in the official Exoplanet Archive (exoplanetarchive.
RONOMY
Bootcamp/Seminar Alum
Python
DOE/NERSC computation
PNAS [2014]](https://image.slidesharecdn.com/pydatabloom16x9-140504231706-phpapp02/85/Data-Science-at-Berkeley-26-320.jpg)
![34 Fitting a straight line to data
0 50 100 150 200 250 300
x
0
100
200
300
400
500
600
700
y
y = 1.33 x + 164
0 50 100 150 200 250 300
x
0
100
200
300
400
500
600
700
y
0.0
1.0
1.0
1.0
0.0
0.0
0.0
0.00.0
0.0
0.0
0.0
0.00.0
0.0
0.0
0.0
0.0
0.0
0.0
Data analysis recipes:
Fitting a model to data⇤
David W. Hogg
Center for Cosmology and Particle Physics, Department of Physics, New York University
Max-Planck-Institut f¨ur Astronomie, Heidelberg
Jo Bovy
Center for Cosmology and Particle Physics, Department of Physics, New York University
Dustin Lang
Department of Computer Science, University of Toronto
Princeton University Observatory
Abstract
We go through the many considerations involved in fitting a model
to data, using as an example the fit of a straight line to a set of points
in a two-dimensional plane. Standard weighted least-squares fitting
is only appropriate when there is a dimension along which the data
points have negligible uncertainties, and another along which all the
uncertainties can be described by Gaussians of known variance; these
D
Fi
Da
Cen
Ma
Jo
Cen
Du
Dep
Prin
tha
dit
line
and
⇤
arXiv:1008.4686v1 [astro-ph.IM] 27 Aug 2010
Sta8s8cal+Inference
Undergraduate$&$Graduate$Training$Mission
Thinking$Data'Literacy$before$
Thinking$Big'Data'Proficiency](https://image.slidesharecdn.com/pydatabloom16x9-140504231706-phpapp02/85/Data-Science-at-Berkeley-29-320.jpg)
Data Science at Berkeley
- 4. diagram'from'Drew'Conway Data$Science$as$a$Discipline • What'are'the'Core$Principles? • Is'it'an'academic$pursuit'to'be'taught'or'a' skillset'to'be'trained? • When'should'they'be'taught?'At'what'level'of' depth/breath? • Who'should'teach'them'and'who'should'know' data'science? • Where'should'investments'be'made?'Does'Data' Science'need'an'intellectual'home'within'insHtuHons?'
- 7. mij = µi + M0j + ↵j log10 (Pi/P0) + E(B V )i ⇥ [RV ⇥ a (1/ i) + b (1/ i)] + ✏ij Bayesian$Distance$Ladder Some-Variable-Stars-show-a-PeriodDBrightness-CorrelaFon i'indexes'over'individual'stars j'indexes'over'wavebands a'and'b'are'fixed'constants'at'each'color Data--134'RR'Lyrae'(ultraviolet'to'infrared) Fit--307-dimensional-model-parameter-inference K'determinisHc'MCMC'model'with'PyMC K'~6'days'for'a'single'run'(one'core) K'parallelism'for'convergence'tests Klein+12;'Klein,JSB+14, “Leavitt Law” HenrieKa-Swan-LeaviK
- 8. Figure 6. Multi-band period–luminosity relations. RRab stars are in blue, RRc stars in red. Blazhko-a↵ected stars are denoted with • Sub-1% distance uncertainty • Precision 3D dust in the Milky Way Bayesian$Distance$Ladder Milky Way projection distance dust Brightness Period Klein+12;'Klein,JSB+14,
- 10. Morphology of LMC RR Lyrae stars 3 the 62 individual science CCDs2 were pro- andard reduction algorithms (bias subtrac- g, etc.) using the computational resources at nergy Research Scientific Computing Center y on individual frames was calibrated with ence sources from the Two Micron All Sky S; Skrutskie et al. 2006) using the astrom- re package (Lang et al. 2010). Photometric performed using same-night observations of Sloan Digital Sky Survey Stripe 82 Standard Ivezi´c et al. 2007). Applying standard cali- dology (e.g., Ofek et al. 2012), we find that a robust scatter in our absolute photometric . 0.02 mag on clear nights (& 50% of the ob- om the Science Verification run). While this be improved with more advanced modeling signatures (e.g., Tucker et al. 2014), we find on is su cient for our scientific objectives. m observing program produced on average Figure 3. V -, I-, and z-band period–magnitude relations (solid lines) derived for the LMC RR Lyrae population, superimposed on scatter plots of the RR Lyrae posteriors (M computed using µPost). The dashed lines denote the 1 prediction intervals for a new RR Lyrae star with known period. used for all of the µi. This standard deviation was selected to be a fractional distance error of 10 per cent (⇡ 5 kpc), which is much larger than the depth of the LMC and significantly larger than (> 2 times) the median posterior µi . To fit the model given by equation 6 ten identical MCMC traces were run, each generating 3.5 million iter- ations. The first 0.5 million were discarded as burn-in and the remaining 3 million were thinned by 300 to result in ten traces of 10,000 iterations each. The Gelman-Rubin conver- now with 15,040 stars... scipy.sparse,'hdf5 Klein..JSB+14 Dust Map basemap 4-✕'improvement'in'distance'error 3D density projection mayavi2
- 13. Our$ML$framework$found$ the$Nearest$Supernova$in$3$ Decades$..‣ Built'&'Deployed'robust,'realKHme' machine'learning'framework,' discovering'>10,000'events'in'>'10' TB'of'imaging' ''''''''→'50+'journal'arHcles ‣ Built'ProbabilisHc'Event' classificaHon'catalogs'with' innovaHve'acHve'learning' hhp://Hmedomain.org hhps://www.nsf.gov/news/news_summ.jsp?cntn_id=122537
- 14. What is the toolbox of the modern (data-driven) scientist? domain training statistics advanced computing database GUI parallel visualization Bayesian machine learning Physics laboratory techniques MCMC MapReduce
- 15. And...How do we teach this with what little time the students have? What is the toolbox of the modern (data-driven) scientist?
- 16. Data-Centric Coursework, Bootcamps, Seminars, & Lecture Series BDAS: Berkeley Data Analytics Stack [Spark, Shark, ...] parallel programming bootcamp ...and entire degree programs
- 17. 2010: 85 campers 2012a: 135 campers Python Bootcamps at Berkeley
- 18. a modern superglue computing language for (data) science ‣ high-level scripting language ‣ open source, huge & growing community in academia & industry ‣ Just in time compilation but also fast numerical computation ‣ Extensive interfaces to 3rd party frameworks
- 19. a modern superglue computing language for (data) science ‣ high-level scripting language ‣ open source, huge & growing community in academia & industry ‣ Just in time compilation but also fast numerical computation ‣ Extensive interfaces to 3rd party frameworks A reasonable lingua franca for scientists...
- 20. 2012b: 210 campers Python Bootcamps at Berkeley 2013a: 253 campers
- 21. ‣ 3 days of live/archive streamed lectures ‣ all open material in GitHub ‣ widely disseminated (e.g., @ NASA) http://pythonbootcamp.info
- 22. Part of the Designated Emphasis in Computational Science & Engineering at Berkeley visualization machine learning database interaction user interface & web frameworks timeseries & numerical computing interfacing to other languages Bayesian inference & MCMC hardware control parallelism
- 23. “Are we alone in the universe? What makes up the missing mass of the universe? ... And maybe the biggest question of all: How in the wide world can you add $3 billion in market capitalization simply by adding .com to the end of a name?” President William Jefferson Clinton Science and Technology Policy Address 21 January 2000 “Add Data Science or Big Data to your course name to increase enrollment by tenfold.” Joshua Bloom Just Now
- 24. Python for Data Science @ Berkeley [Sept 2013]
- 25. 64% 36% female male 8% 4% 8% 12% 4% 12% 8% 16% 16% 12%Psychology Astronomy Neuroscience Biostatistics Physics Chemical Engineering ISchool Earth and Planetary Sciences Industrial Engineering Mechanical Engineering “Parallel Image Reconstruction from Radio Interferometry Data” “Graph Theory Analysis of Growing Graphs” http://mb3152.github.io/Graph-Growth/ “Realtime Prediction of Activity Behavior from Smartphone” “Bus Arrival Time Prediction in Spain”
- 26. Time domain preprocessing - Start with raw photometry! - Gaussian process detrending! - Calibration! - Petigura & Marcy 2012! ! Transit search - Matched filter! - Similar to BLS algorithm (Kovcas+ 2002)! - Leverages Fast-Folding Algorithm O(N^2) → O(N log N) (Staelin+ 1968)! ! Data validation - Significant peaks in periodogram, but inconsistent with exoplanet transit TERRA – optimized for small planets Detrended/calibrated photometry TERRA RawFlux(ppt)CalibratedFlux Erik Petigura Berkeley Astro Grad Student Prevalence of Earth-size planets orbiting Sun-like stars Erik A. Petiguraa,b,1 , Andrew W. Howardb , and Geoffrey W. Marcya a Astronomy Department, University of California, Berkeley, CA 94720; and b Institute for Astronomy, University of Hawaii at Manoa, Honolulu, HI 96822 Contributed by Geoffrey W. Marcy, October 22, 2013 (sent for review October 18, 2013) Determining whether Earth-like planets are common or rare looms as a touchstone in the question of life in the universe. We searched for Earth-size planets that cross in front of their host stars by examining the brightness measurements of 42,000 stars from National Aeronautics and Space Administration’s Kepler mission. We found 603 planets, including 10 that are Earth size (1 − 2 R⊕) and receive comparable levels of stellar energy to that of Earth (0:25 − 4 F⊕). We account for Kepler’s imperfect detectability of such planets by injecting synthetic planet–caused dimmings into the Kepler brightness measurements and recording the fraction detected. We find that 11 ± 4% of Sun-like stars harbor an Earth- size planet receiving between one and four times the stellar inten- sity as Earth. We also find that the occurrence of Earth-size planets is constant with increasing orbital period (P), within equal intervals of logP up to ∼200 d. Extrapolating, one finds 5:7+1:7 −2:2 % of Sun-like stars harbor an Earth-size planet with orbital periods of 200–400 d. extrasolar planets | astrobiology The National Aeronautics and Space Administration’s (NASA’s) Kepler mission was launched in 2009 to search for planets that transit (cross in front of) their host stars (1–4). The resulting dimming of the host stars is detectable by measuring their bright- ness, and Kepler monitored the brightness of 150,000 stars every 30 min for 4 y. To date, this exoplanet survey has detected more than 3,000 planet candidates (4). We searched for transiting planets in Kepler brightness mea- surements using our custom-built TERRA software package described in previous works (6, 9) and in SI Appendix. In brief, TERRA conditions Kepler photometry in the time domain, re- moving outliers, long timescale variability (>10 d), and systematic errors common to a large number of stars. TERRA then searches for transit signals by evaluating the signal-to-noise ratio (SNR) of prospective transits over a finely spaced 3D grid of orbital period, P, time of transit, t0, and transit duration, ΔT. This grid-based search extends over the orbital period range of 0.5–400 d. TERRA produced a list of “threshold crossing events” (TCEs) that meet the key criterion of a photometric dimming SNR ratio SNR > 12. Unfortunately, an unwieldy 16,227 TCEs met this cri- terion, many of which are inconsistent with the periodic dimming profile from a true transiting planet. Further vetting was performed by automatically assessing which light curves were consistent with theoretical models of transiting planets (10). We also visually inspected each TCE light curve, retaining only those exhibiting a consistent, periodic, box-shaped dimming, and rejecting those caused by single epoch outliers, correlated noise, and other data anomalies. The vetting process was applied homogeneously to all TCEs and is described in further detail in SI Appendix. To assess our vetting accuracy, we evaluated the 235 Kepler objects of interest (KOIs) among Best42k stars having P > 50 d, which had been found by the Kepler Project and identified as planet candidates in the official Exoplanet Archive (exoplanetarchive. RONOMY Bootcamp/Seminar Alum Python DOE/NERSC computation PNAS [2014]
- 29. 34 Fitting a straight line to data 0 50 100 150 200 250 300 x 0 100 200 300 400 500 600 700 y y = 1.33 x + 164 0 50 100 150 200 250 300 x 0 100 200 300 400 500 600 700 y 0.0 1.0 1.0 1.0 0.0 0.0 0.0 0.00.0 0.0 0.0 0.0 0.00.0 0.0 0.0 0.0 0.0 0.0 0.0 Data analysis recipes: Fitting a model to data⇤ David W. Hogg Center for Cosmology and Particle Physics, Department of Physics, New York University Max-Planck-Institut f¨ur Astronomie, Heidelberg Jo Bovy Center for Cosmology and Particle Physics, Department of Physics, New York University Dustin Lang Department of Computer Science, University of Toronto Princeton University Observatory Abstract We go through the many considerations involved in fitting a model to data, using as an example the fit of a straight line to a set of points in a two-dimensional plane. Standard weighted least-squares fitting is only appropriate when there is a dimension along which the data points have negligible uncertainties, and another along which all the uncertainties can be described by Gaussians of known variance; these D Fi Da Cen Ma Jo Cen Du Dep Prin tha dit line and ⇤ arXiv:1008.4686v1 [astro-ph.IM] 27 Aug 2010 Sta8s8cal+Inference Undergraduate$&$Graduate$Training$Mission Thinking$Data'Literacy$before$ Thinking$Big'Data'Proficiency
- 30. Versioning+&+Reproducibility “Recently, the scientific community was shaken by reports that a troubling proportion of peer-reviewed preclinical studies are not reproducible.” McNutt, 2014 http://www.sciencemag.org/content/343/6168/229.summary K'Git'has'emerged'as'the'de'facto'versioning'tool K'Berkeley'Common'Environment'(BCE)'Sonware'Stack K'“Reproducible'and'CollaboraHve'StaHsHcal'Data'Science”'(StaHsHcs'157:'P.'Stark) Undergraduate$&$Graduate$Training$Mission Thinking$Data'Literacy$before$ Thinking$Big'Data'Proficiency
- 32. Berkeley Data Analytics Stack A Comprehensive Big Data Reference Architecture AMP! Alpha or Soon! AMP! Released! BSD/Apache! 3rd Party Open Source! Apache Mesos! YARN Resource Manager! Resource! HDFS / Hadoop Storage! Tachyon! Storage! ! Apache Spark! Spark Streaming! ML-lib! Processing! and Data ! Management! Applications: Traffic, Carat, Genomics, 3rd Party! Tools: Visualization, Data Cleaning, …! Shark (SQL)! BlinkDB! GraphX! MLBase! Analytics! Frameworks! Spark-R! Berkeley$Data$Analy;cs$Stack A,Comprehensive,Big,Data,Reference, Architecture Methodology'innovaHon:'InvenHng'What’s'Next'in'Big'Data'AnalyHcs Testing the Vision/Early Adopters/Momentum 2010+
- 35. Created by Natalia Bilenko. Data source: PubMed Central. http://sciencereview.berkeley.edu/bsr_design/Issue26/datascience/cover/
- 37. Berkeley Institute for Data Science http://bitly.com/bundles/fperezorg/1 “Bold new partnership launches to harness potential of data scientists and big data” Founded'in'December'2013'as'a'result'of'a'year+'long'naHonal'selecHon'process $37.8M'over'5'years,'along'with'University'of'Washington'&'NYU ‣ An'accelerator'for'dataKdriven'discovery ‣ An'agent$of$change'in'the'modern'university'as'Data'Science'takes'hold ‣ An'incubator'for'the'next'generaHon'of'Data'Science'technology'and'pracHce
- 39. BIDS goals ‣ Support$meaningful$and$sustained$interac;ons$and$collabora;ons$ between'Methodology'fields'&'Science'domains'to'recognize'what'it'takes'to' move'these'fields'forward ‣ Establish$new$Data$Science$career$paths$that$are$longAterm$and$sustainable • A'generaHon'of'mulHKdisciplinary'scienHsts'in'dataKintensive'science • A'generaHon'of'data'scienHsts'focused'on'tool'development ‣ Build$an$ecosystem$of$analy;cal$tools,$teaching,$&$research$prac;ces • Sustainable,'reusable,'extensible,'easy'to'learn'and'to'translate'across' research'domains • Enables'scienHsts'to'spend'more'Hme'focusing'on'their'science 37
- 42. Berkeley Institute for Data Science Opening
- 43. Berkeley Institute for Data Science Opening