Lock in $30 Savings on PRO—Offer Ends Soon! ⏳

Data Science for the STEM undergrad

Data Science for the STEM undergrad

Lightning talk I gave at PyData 2017

Avatar for jehan60188

jehan60188

July 07, 2017
Tweet

Other Decks in Education

Transcript

  1. Agenda • About me • What is data science and

    machine learning (DSML) • Benefits of learning DSML • Benefits of proficiency in python libraries • *ATLAB s*cks • Why STEM needs DSML • Why STEM students are best positioned for DSML education • Basic techniques and best practices • Python libraries
  2. What is Data Science? •5 types of questions •Is this

    A/B (classification) •Is this weird (anomaly detection) © Wikimedia commons
  3. What is Data Science? •5 types of questions •Is this

    A/B (classification) •Is this weird (anomaly detection) •How much (regression) © Wikimedia commons
  4. What is Data Science? •5 types of questions •Is this

    A/B (classification) •Is this weird (anomaly detection) •How much (regression) •How are the data organized (clustering) © Wikimedia commons
  5. What is Data Science? •5 types of questions •Is this

    A/B (classification) •Is this weird (anomaly detection) •How much (regression) •How are the data organized (clustering) •What next (reinforcement) © Wikimedia commons
  6. What is machine learning? •Programming Inputs Program Computer Outputs Inputs

    Outputs Computer Program •Machine learning NEW inputs! NEW OUTPUTS!
  7. Why learn DSML • Everywhere • Face (emotion!) detection ©

    Steve Jurvetson: https://www.flickr.com/photos/jurvetson/6866541534
  8. Why learn DSML • Everywhere • Face (emotion!) detection •

    Self driving cars © Epety ET: https://www.youtube.com/watch?v=kMMbW96nMW8
  9. Why learn DSML • Everywhere • Face (emotion!) detection •

    Self driving cars • Weather forecasting © NOAA https://www.ncdc.noaa.gov/news/historical-probability-of-white-christmas
  10. Why learn DSML • Everywhere • Face (emotion!) detection •

    Self driving cars • Weather forecasting • Easy • Calculus • Statistics and probability • Linear algebra • Programming © Randall Munroe: https://xkcd.com/1838/
  11. Why learn machine learning • Everywhere • Face (emotion!) detection

    • Self driving cars • Weather forecasting • Easy • Calculus • Statistics • Linear algebra • Programming • Python • scikit-learn © Randall Munroe: https://xkcd.com/353/
  12. Python is awesome import sklearn.linear_model as LM X = np.array([1,2,3,4,5]).reshape(-1,1)

    X = X.astype(float) Y = np.array([1,3,3,3,5]).reshape(-1,1) Y = Y.astype(float) m_model = LM.LinearRegression() m_model.fit(X,Y) print( 'Y = ' +str( m_model.coef_[0][0]) + '*X+’ +\ str(m_model.intercept_[0])) >>Y = 0.8*X+0.6
  13. No other ML package is as succinct and well thought

    out a sci-kit learn Jake VanderPlas, PyData 2017
  14. How do we proceed? • Students • Petition your school

    for a data science minor • Check your statistics department • Choose DSML-oriented solutions • Educators • Introduce programming early • Create DSML-oriented examples/solutions • Practitioners • Engage highschool/undergrad students/interns • Offer adult education seminars at your library • Get involved in politics! • Vote/volunteer for STEM-positive persons • RUN on a STEM-positive platform!