Upgrade to Pro — share decks privately, control downloads, hide ads and more …

The art and science of teaching data science (UCL)

The art and science of teaching data science (UCL)

Modern statistics is fundamentally a computational discipline, but too often this fact is not reflected in our statistics curricula. With the rise of data science it has become increasingly clear that students want, expect, and need explicit training in this area of the discipline. Additionally, recent curricular guidelines clearly state that working with data requires extensive computing skills and that statistics students should be fluent in accessing, manipulating, analyzing, and modeling with professional statistical analysis software. In this talk, we introduce the design philosophy behind an introductory data science course, discuss in progress and future research on student learning as well as new directions in assessment and tooling as we scale up the course.

Mine Cetinkaya-Rundel

February 27, 2020
Tweet

More Decks by Mine Cetinkaya-Rundel

Other Decks in Education

Transcript

  1. Image credit: Thomas Pedersen, data-imaginist.com/art
    the art and science
    of teaching data science
    mine çetinkaya-rundel
    mine-cetinkaya-rundel
    [email protected]
    @minebocek
    university of edinburgh
    bit.ly/art-sci-ucl

    View full-size slide

  2. 2016 GAISE
    1. Teach statistical thinking.
    ‣ Teach statistics as an investigative process of problem-solving and decision making.
    Students should not leave their introductory statistics course with the mistaken impression
    that statistics consists of an unrelated collection of formulas and methods. Rather,
    students should understand that statistics is a problem-solving and decision making
    process that is fundamental to scientific inquiry and essential for making sound decisions.
    ‣ Give students experience with multivariable thinking. We live in a complex world in
    which the answer to a question often depends on many factors. Students will encounter
    such situations within their own fields of study and everyday lives. We must prepare our
    students to answer challenging questions that require them to investigate and explore
    relationships among many variables. Doing so will help them to appreciate the value of
    statistical thinking and methods.
    2. Focus on conceptual understanding.
    3. Integrate real data with a context and purpose.
    4. Foster active learning.
    5. Use technology to explore concepts and analyse data.
    6. Use assessments to improve and evaluate student learning.
    amstat.org/asa/files/pdfs/GAISE/GaiseCollege_Full.pdf

    View full-size slide

  3. 2016 GAISE
    1. Teach statistical thinking.
    ‣ Teach statistics as an investigative process of problem-solving and decision making.
    Students should not leave their introductory statistics course with the mistaken impression
    that statistics consists of an unrelated collection of formulas and methods. Rather,
    students should understand that statistics is a problem-solving and decision making
    process that is fundamental to scientific inquiry and essential for making sound decisions.
    ‣ Give students experience with multivariable thinking. We live in a complex world in
    which the answer to a question often depends on many factors. Students will encounter
    such situations within their own fields of study and everyday lives. We must prepare our
    students to answer challenging questions that require them to investigate and explore
    relationships among many variables. Doing so will help them to appreciate the value of
    statistical thinking and methods.
    2. Focus on conceptual understanding.
    3. Integrate real data with a context and purpose.
    4. Foster active learning.
    5. Use technology to explore concepts and analyse data.
    6. Use assessments to improve and evaluate student learning.
    amstat.org/asa/files/pdfs/GAISE/GaiseCollege_Full.pdf
    1 NOT a
    commonly used
    subset of tests
    and intervals
    and produce
    them with hand
    calculations

    View full-size slide

  4. 2016 GAISE
    1. Teach statistical thinking.
    ‣ Teach statistics as an investigative process of problem-solving and decision making.
    Students should not leave their introductory statistics course with the mistaken impression
    that statistics consists of an unrelated collection of formulas and methods. Rather,
    students should understand that statistics is a problem-solving and decision making
    process that is fundamental to scientific inquiry and essential for making sound decisions.
    ‣ Give students experience with multivariable thinking. We live in a complex world in
    which the answer to a question often depends on many factors. Students will encounter
    such situations within their own fields of study and everyday lives. We must prepare our
    students to answer challenging questions that require them to investigate and explore
    relationships among many variables. Doing so will help them to appreciate the value of
    statistical thinking and methods.
    2. Focus on conceptual understanding.
    3. Integrate real data with a context and purpose.
    4. Foster active learning.
    5. Use technology to explore concepts and analyse data.
    6. Use assessments to improve and evaluate student learning.
    amstat.org/asa/files/pdfs/GAISE/GaiseCollege_Full.pdf
    2 Multivariate
    analysis
    requires the use
    of computing

    View full-size slide

  5. 2016 GAISE
    1. Teach statistical thinking.
    ‣ Teach statistics as an investigative process of problem-solving and decision making.
    Students should not leave their introductory statistics course with the mistaken impression
    that statistics consists of an unrelated collection of formulas and methods. Rather,
    students should understand that statistics is a problem-solving and decision making
    process that is fundamental to scientific inquiry and essential for making sound decisions.
    ‣ Give students experience with multivariable thinking. We live in a complex world in
    which the answer to a question often depends on many factors. Students will encounter
    such situations within their own fields of study and everyday lives. We must prepare our
    students to answer challenging questions that require them to investigate and explore
    relationships among many variables. Doing so will help them to appreciate the value of
    statistical thinking and methods.
    2. Focus on conceptual understanding.
    3. Integrate real data with a context and purpose.
    4. Foster active learning.
    5. Use technology to explore concepts and analyse data.
    6. Use assessments to improve and evaluate student learning.
    amstat.org/asa/files/pdfs/GAISE/GaiseCollege_Full.pdf
    3 NOT use
    technology that
    is only
    applicable in the
    intro course or
    that doesn’t
    follow good
    science
    principles

    View full-size slide

  6. 2016 GAISE
    1. Teach statistical thinking.
    ‣ Teach statistics as an investigative process of problem-solving and decision making.
    Students should not leave their introductory statistics course with the mistaken impression
    that statistics consists of an unrelated collection of formulas and methods. Rather,
    students should understand that statistics is a problem-solving and decision making
    process that is fundamental to scientific inquiry and essential for making sound decisions.
    ‣ Give students experience with multivariable thinking. We live in a complex world in
    which the answer to a question often depends on many factors. Students will encounter
    such situations within their own fields of study and everyday lives. We must prepare our
    students to answer challenging questions that require them to investigate and explore
    relationships among many variables. Doing so will help them to appreciate the value of
    statistical thinking and methods.
    2. Focus on conceptual understanding.
    3. Integrate real data with a context and purpose.
    4. Foster active learning.
    5. Use technology to explore concepts and analyse data.
    6. Use assessments to improve and evaluate student learning.
    amstat.org/asa/files/pdfs/GAISE/GaiseCollege_Full.pdf
    4 Data analysis
    isn’t just
    inference and
    modelling, it’s
    also data
    importing,
    cleaning,
    preparation,
    exploration, and
    visualisation

    View full-size slide

  7. a course that satisfies these four
    points is looking more like today’s
    intro data science courses than
    (most) intro stats courses
    but this is not because
    intro stats is inherently
    “bad for you”
    instead it is because it’s time to visit
    intro stats in light of emergence of
    data science

    View full-size slide

  8. fundamentals of
    data & data viz,
    confounding variables,
    Simpson’s paradox
    +
    R / RStudio,
    R Markdown, simple Git
    tidy data, data frames
    vs. summary tables,
    recoding &
    transforming,
    web scraping & iteration
    +
    collaboration on GitHub

    View full-size slide

  9. fundamentals of
    data & data viz,
    confounding variables,
    Simpson’s paradox
    +
    R / RStudio,
    R Markdown, simple Git
    tidy data, data frames
    vs. summary tables,
    recoding &
    transforming,
    web scraping & iteration
    +
    collaboration on GitHub
    building & selecting
    models,
    visualising interactions,
    prediction & validation,
    inference via simulation

    View full-size slide

  10. fundamentals of
    data & data viz,
    confounding variables,
    Simpson’s paradox
    +
    R / RStudio,
    R Markdown, simple Git
    tidy data, data frames
    vs. summary tables,
    recoding &
    transforming,
    web scraping & iteration
    +
    collaboration on GitHub
    building & selecting
    models,
    visualising interactions,
    prediction & validation,
    inference via simulation
    data science ethics,
    text analysis,
    Bayesian inference
    +
    communication &
    dissemination

    View full-size slide

  11. three questions that keep me up at night…
    1 what should students learn?
    2 how will students learn best?
    3 what tools will enhance student learning?

    View full-size slide

  12. three questions that keep me up at night…
    1 what should students learn?
    2 how will students learn best?
    3 what tools will enhance student learning?
    content
    pedagogy
    infrastructure

    View full-size slide

  13. ex. 1
    money in politics

    View full-size slide

  14. ✴ web scraping
    ✴ text parsing
    ✴ data types
    ✴ regular expressions

    View full-size slide

  15. ✴ web scraping
    ✴ text parsing
    ✴ data types
    ✴ regular expressions
    ✴ iteration

    View full-size slide

  16. ✴ web scraping
    ✴ text parsing
    ✴ data types
    ✴ regular expressions
    ✴ iteration
    ✴ data visualisation
    ✴ interpretation

    View full-size slide

  17. ✴ web scraping
    ✴ text parsing
    ✴ data types
    ✴ regular expressions
    ✴ iteration
    ✴ data visualisation
    ✴ interpretation
    ✴ data science ethics

    View full-size slide

  18. Project: The North South Divide: University Edition
    Question: Does the geographical location of a UK university affect its
    university score?
    Team: Fried Egg Jelly Fish

    View full-size slide

  19. ex. 2
    fisheries of the world

    View full-size slide

  20. ✴ data joins

    View full-size slide

  21. ✴ data joins
    ✴ data science ethics

    View full-size slide

  22. ✴ data joins
    ✴ data science ethics
    ✴ critique
    ✴ improving data
    visualisations

    View full-size slide

  23. ✴ data joins
    ✴ data science ethics
    ✴ critique
    ✴ improving data
    visualisations
    ✴ mapping

    View full-size slide

  24. Project: 2016 US Election Redux
    Question: Would the outcome of the 2016 US Presidential Elections been
    different had Bernie Sanders been the Democrat candidate?
    Team: 4 Squared

    View full-size slide

  25. ex. 3
    spam filters

    View full-size slide

  26. ✴ logistic regression
    ✴ prediction

    View full-size slide

  27. ✴ logistic regression
    ✴ prediction
    ✴ decision errors
    ✴ sensitivity /
    specificity
    ✴ intuition around
    loss functions

    View full-size slide

  28. Project: Spotify Top 100 Tracks of 2017/18
    Question: Is it possible to predict the year a song made the Top Tracks
    playlist based on its metadata?
    Team: weR20
    year ~ danceability + energy + key + loudness + mode + speechiness +
    acousticness + instrumentalness + liveness + valence + tempo +
    duration_s
    2017
    name artists
    I'm the One DJ Khaled
    Redbone Childish Gambino
    Sign of the Times Harry Styles
    2018
    name artists
    Everybody Dies In Their Nightmares XXXTENTACION
    Jocelyn Flores XXXTENTACION
    Plug Walk Rich The Kid
    Moonlight XXXTENTACION
    Nevermind Dennis Lloyd
    In My Mind Dynoro
    changes XXXTENTACION

    View full-size slide

  29. teams: weekly labs in teams +
    periodic team evaluations +
    term project in teams
    peer feedback: used
    minimally so far, but
    positive experience
    “minute paper”: weekly online
    quizzes ending with a brief
    reflection of the week’s material

    View full-size slide

  30. teams: weekly labs in teams +
    periodic team evaluations +
    term project in teams
    peer feedback: used
    minimally so far, but
    positive experience
    “minute paper”: weekly online
    quizzes ending with a brief
    reflection of the week’s material
    creativity: assignments that
    make room for creativity

    View full-size slide

  31. infrastructure

    View full-size slide

  32. three questions that keep me up at night…
    1 what should students learn?
    2 how will students learn best?
    3 what tools will enhance student learning?
    content
    pedagogy
    infrastructure
    four
    4 how can we assess any of this?
    assessment

    View full-size slide

  33. data: 205 open-ended student projects
    over 4 years
    group 1:
    learned R & intro
    statistics using
    base R
    group 2:
    learned R & intro
    statistics using
    tidyverse*
    * starting before the term
    tidyverse was coined.
    same assignment, same(ish) dataset
    measures: creativity, depth and the
    complexity of multivariate visualisations
    in progress: retrospective study

    View full-size slide

  34. depth
    - consistent theme
    throughout the project
    - relevant data for each
    analysis
    0
    20
    40
    60
    0 1 2
    Depth score
    Number of projects
    Syntax
    Base R
    Tidyverse
    Depth scores by syntax

    View full-size slide

  35. 0
    20
    40
    0 1 2 3 4
    Creativity score
    Number of projects
    Syntax
    Base R
    Tidyverse
    Creativity scores by syntax
    creativity
    - creation of new variables
    - transformation of existing
    variables
    - subgroup analysis
    - use of a subset of data for
    the entire project

    View full-size slide

  36. 0
    25
    50
    75
    0 1 2
    Multivariate visualisation score
    Number of projects
    Syntax
    Base R
    Tidyverse
    Multivariate visualisation by syntax
    multivariate visualisation
    - visualisation with 3+ variables
    - effective interpretations of
    visualisations

    View full-size slide

  37. planned: longitudinal study
    motivation: higher conversion rate to stat 2
    explorations:
    retention, especially of
    students from under-
    represented
    backgrounds
    preparation and
    confidence for applied
    and collaborative
    projects

    View full-size slide

  38. Image credit:
    Thomas Pedersen, data-imaginist.com/art
    the art and science
    of teaching data science
    mine çetinkaya-rundel
    mine-cetinkaya-rundel
    [email protected]
    @minebocek
    university of edinburgh
    bit.ly/art-sci-ucl

    View full-size slide