Video e x ercises : Slides & transcript C OU R SE C R E ATION - PowerPoint PPT Presentation

Video e x ercises : Slides & transcript C OU R SE C R E ATION AT DATAC AMP Hadrien Lacroi x Content De v eloper

Video e x ercises First part of a lesson Learning objecti v e Presentation COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

Str u ct u red scripts E x act w ords recorded What → Wh y → Ho w Sentences sho u ld � o w nat u rall y Be brief ( ideall y 400-500 w ords ) Ma x 600 w ords COURSE CREATION AT DATACAMP

B u ild a narrati v e Stor y telling Be ADEPT 1. Use an Analog y 2. Dra w a Diagram 3. Pro v ide a concrete E x ample 4. Describe it in Plain English 5. Pro v ide a Technical de � nition COURSE CREATION AT DATACAMP

Timeless scripts Don ' t Do Manip u lating Time Series Data in P y thon co u rse on time series manip u lation in P y thon I ' m a Data Scientist at Data Compan y X I ' m a Data Scientist I ' m w riting a book on Machine Learning Follo w me on T w i � er for u pdates The c u rrent v ersion of this package ... < Foc u s on the u tilit y, not the implementation > COURSE CREATION AT DATACAMP

D y namic slides D y namic : Add animations to lists Add animations to code A v oid " deadtime " Add mo v ement ( at least ) e v er y 30 seconds Animate b u llet points COURSE CREATION AT DATACAMP

B u llet points : What y o u sho u ld NOT do Machine learning w ork � o w Data e x ploration Data processing Modeling E v al u ation Impro v ements COURSE CREATION AT DATACAMP

B u llet points : What y o u sho u ld do Machine Learning w ork � o w Machine Learning w ork � o w 1. Data e x ploration {{1}} → 1. Data e x ploration 2. Data processing {{2}} → 2. Data processing 3. Modeling {{3}} → 3. Modeling 4. E v al u ation {{4}} → 4. E v al u ation 5. Impro v ements {{5}} → 5. Impro v ements COURSE CREATION AT DATACAMP

Code : What y o u sho u ld NOT do # Define trainControl function fitControl <- trainControl(method = "adaptive_cv", number = 3, repeats = 3, adaptive = list(min = 3, alpha = 0.05, method = "BT", complete = FALSE), search = "random") # Start timer tic() # Train model svm_model_voters_ar <- train(turnout16_2016 ~ ., data = voters_train_data, method = "svmPoly", trControl = fitControl, COURSE CREATION AT DATACAMP

Code : What y o u sho u ld do # Define trainControl function fitControl <- trainControl(method = "adaptive_cv", number = 3, repeats = 3, adaptive = list(min = 3, alpha = 0.05, method = "BT", complete = FALSE), search = "random") # Start timer tic() # Train model svm_model_voters_ar <- train(turnout16_2016 ~ ., data = voters_train_data, method = "svmPoly", trControl = fitControl, verbose = FALSE, tuneLength = 6) # Stop timer toc() COURSE CREATION AT DATACAMP

S y nta x highlighting COURSE CREATION AT DATACAMP

S y nta x highlighting Inp u t onl y : { p y thon }, { r } COURSE CREATION AT DATACAMP

Images : B u ilding a diagram (1) COURSE CREATION AT DATACAMP

Images : B u ilding a diagram (2) COURSE CREATION AT DATACAMP

Images : B u ilding a diagram (3) COURSE CREATION AT DATACAMP

Images : B u ilding a diagram (4) COURSE CREATION AT DATACAMP

Images : B u ilding a diagram (5) COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

Let ' s practice ! C OU R SE C R E ATION AT DATAC AMP

Interacti v e e x ercises C OU R SE C R E ATION AT DATAC AMP Sara Billen Content De v eloper

E x ercise t y pes M u ltiple Choice e x ercise M u ltiple Choice w ith Console e x ercise Coding e x ercise Iterati v e e x ercise Seq u ential e x ercise COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

Writing good sample code Sample code Sol u tion code # Open a file: file # Open a file: file file = open(____, file = open('moby_dick.txt, mode='____') mode='r') # Print it # Print it ____ print(file.read()) # Check whether file is closed # Check whether file is closed ____(file.closed) print(file.closed) # Close file # Close file () () COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

Let ’ s start e x ercising ! C OU R SE C R E ATION AT DATAC AMP

Respect the g u idelines C OU R SE C R E ATION AT DATAC AMP Shon Ino uy e Content Q u alit y Anal y st

Objecti v e of content g u idelines Goal : Enable st u dents to learn data science in a f u n and engaging w a y Content G u idelines : Ens u re that all content w e de v elop is in s y nc w ith o u r goal Pro v ide u sers w ith the best possible u ser e x perience COURSE CREATION AT DATACAMP

Reasoning : Consistenc y DataCamp has a rapidl y e x panding u ser base and content librar y Consistenc y across content to maintain prod u ct v ision in a scalable fashion COURSE CREATION AT DATACAMP

Reasoning : Optimi z ation for engagement Data collection on st u dent interactions w ith DataCamp allo w s u s to e x plore : Completion rate Percentage of st u dents asking for hints / sol u tions Co u rse rating Insight from pedagogical ( teaching - related ) research MIT st u d y recommending v ideo length to be < 6 min u tes COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

Instr u ction % Asked % Asked E x ercise Completion % of Length Hint Sol u tion Rate E x ercises (0, 100] 0.08 0.05 0.9971 5 (100, 200] 0.14 0.10 0.9926 16 (200, 300] 0.17 0.13 0.9922 21 (300, 400] 0.21 0.16 0.9898 20 (400, 500] 0.24 0.18 0.9889 15 (500, 600] 0.26 0.20 0.9886 10 Dri v es completion rates and % st u dents asking for hints / sol u tions 60 (0 9971) 99 71% i l ti 84% l ti COURSE CREATION AT DATACAMP

Other g u idelines G u ideline Reasoning Aim for co u rses to be completable w ithin 4 Shorter co u rses are more digestible and ho u rs ha v e higher completion rates Limit e x ercises to 15 lines of sample / sol u tion Ens u res that instr u ctions � t w ell w ithin a code page and limits scrolling Limit the n u mber of m u ltiple choice q u estions Ens u res learners do most of their learning in a co u rse to no more than 5 b y doing COURSE CREATION AT DATACAMP

Reso u rces Content G u idelines Co u rse Editor feat u res for Content G u idelines COURSE CREATION AT DATACAMP

Let ' s practice ! C OU R SE C R E ATION AT DATAC AMP

GitH u b for co u rse re v ie w C OU R SE C R E ATION AT DATAC AMP Jeroen Hermans Head of Content Engineering

Back to the repo GitH u b G u ides COURSE CREATION AT DATACAMP

P u ll req u ests ( PR ) COURSE CREATION AT DATACAMP

P u ll req u ests ( PR ) COURSE CREATION AT DATACAMP

P u ll req u ests ( PR ) COURSE CREATION AT DATACAMP

P u ll req u ests ( PR ) COURSE CREATION AT DATACAMP

Merge conflicts COURSE CREATION AT DATACAMP

Diffs COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

Re v ie w comments COURSE CREATION AT DATACAMP

Resol v e comments COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

COURSE CREATION AT DATACAMP

Let ' s practice ! C OU R SE C R E ATION AT DATAC AMP