Info Session — Mentor-Led Data Science & AI Program

Register
Academy

Module 12

Big data with Spark

Spark DataFrames, distributed preprocessing, feature engineering at scale, model training, tuning with k-fold cross-validation and grid search.

Start lesson 1 Take the quiz9 lessons · 72 min · 6 quiz questions

Outcome

What you will be able to do

You can build and tune a Spark ML pipeline on data that does not fit on one machine.

Lessons

Work through these in order

  1. 01Spark Crash CourseThe dataset contains bike rental info from 2011 and 2012 in the Capital bikeshare system, plus additional relevant information such as weather. 8 min
  2. 02DataPreProcessingcat > movies.csv <<EOF name,rating,studio,date Avengers Endgame, 5, marvel,1260759144 Batman Vs Superman,4 ,DC,835355664 The Joker ,, DC,835355681 Frozen,4,Disney,835355604 Hitman,,, EOF %sh ls -alh … 8 min
  3. 03MLeap-Orange Telecom Customer Churnfrom pyspark.ml.feature import VectorAssembler from pyspark.ml.classification import LogisticRegression from pyspark.ml import Pipeline from pyspark.sql.functions import udf ​from pyspark.sql.types… 8 min
  4. 04Capital Bike Rental-Python-1How to Ingest the data into Spark DataFrame. How to clean the Data with DataFrame, SQL Query. How to create a Machine Learning Pipeline. How to train a Machine Learning model. How to save & read the… 8 min
  5. 05Data IngestionWe begin by loading our data, which is stored in the CSV format. 8 min
  6. 06Data UnderstandingNow that we have preprocessed our features, we can quickly visualize our data to get a sense of whether the features are meaningful. 8 min
  7. 07Data Processing (Feature Engineering)For each of the categorical columns, we are going to create one StringIndexer where we 8 min
  8. 08Train the modelRandom forests and ensembles of decision trees are more powerful than a single decision tree alone. 8 min
  9. 09Model tuning: Preparing K-fold Cross Validation and Grid Searchfor best model selection and makes sure that there's no overfitting. 8 min

Assessment

Module quiz — 70% to pass

6 questions mixing concept checks and short code-output problems. Graded on the server with per-question explanations afterwards, unlimited retakes, and a badge with a verification code the moment you pass.