Scribd
Upload
7 views

DA Lab Program-6

Uploaded by

Diksha Padiyar
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
7 views

DA Lab Program-6

Uploaded by

Diksha Padiyar
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

DATA ANALYTICS LABORATORY (21CSL66)

6. Utilize Apache Spark to achieve the following tasks for the given dataset,

a) Find the movie with the lowest average rating with RDD.

b) Identify users who have rated the most movies.

c) Explore the distribution of ratings over time.

d)Find the highest-rated movies with a minimum number of ratings.

Steps to be followed:

Prerequisites:

• Apache Spark Setup: Ensure you have Apache Spark installed and configured.

(If not go to terminal and install pyspark using pip install pyspark)

• Dataset: Load the dataset.

o movies dataset has columns: movie_id, title, genres.

o ratings dataset has columns: user_id, movie_id, rating.

• Loading Data: First, load the data into Spark RDDs or DataFrames.

from pyspark.sql import SparkSession

# Create RDDs

#movies_rdd = movies_df.rdd

#ratings_rdd = ratings_df.rdd

# Initialize Spark session

spark = SparkSession.builder.appName("MovieRatingsAnalysis").getOrCreate()

1
# Load datasets

movies_df = spark.read.csv("/Users/amithpradhaan/Desktop/ml-latest-small/movies.csv",
header=True, inferSchema=True)

ratings_df = spark.read.csv("/Users/amithpradhaan/Desktop/ml-latest-small/ratings.csv",
header=True, inferSchema=True)

# Create RDDs

movies_rdd = movies_df.rdd

ratings_rdd = ratings_df.rdd

a) Find the Movie with the Lowest Average Rating Using RDD.

# Compute average ratings

avg_ratings_rdd = ratings_rdd.map(lambda x: (x['movieId'], (x['rating'], 1))) \

.reduceByKey(lambda x, y: (x[0] + y[0], x[1] + y[1])) \

.mapValues(lambda x: x[0] / x[1])

# Find the movie with the lowest average rating

lowest_avg_rating = avg_ratings_rdd.sortBy(lambda x: x[1]).first()

print(f"Movie with the lowest average rating: {lowest_avg_rating}")

b) Identify Users Who Have Rated the Most Movies.

# Compute number of ratings per user

user_ratings_count = ratings_rdd.map(lambda x: (x['userId'], 1)) \

.reduceByKey(lambda x, y: x + y) \

.sortBy(lambda x: x[1], ascending=False)

# Get top users

top_users = user_ratings_count.take(10)
2
print(f"Top users by number of ratings: {top_users}")

c) Explore the Distribution of Ratings Over Time.

from pyspark.sql.functions import from_unixtime, year, month

# Convert timestamp to date and extract year and month

ratings_df = ratings_df.withColumn("year", year(from_unixtime(ratings_df['timestamp']))) \

.withColumn("month", month(from_unixtime(ratings_df['timestamp'])))

# Group by year and month to get rating counts

ratings_over_time = ratings_df.groupBy("year", "month").count().orderBy("year", "month")

# Show distribution

ratings_over_time.show()

d) Find the Highest-Rated Movies with a Minimum Number of Ratings.

Set a minimum number of ratings, ex: 100.

# Compute average ratings and count ratings per movie

movie_ratings_stats = ratings_rdd.map(lambda x: (x['movieId'], (x['rating'], 1))) \

.reduceByKey(lambda x, y: (x[0] + y[0], x[1] + y[1])) \

.mapValues(lambda x: (x[0] / x[1], x[1])) # (avg_rating, count)

# Filter movies with a minimum number of ratings

min_ratings = 100

qualified_movies = movie_ratings_stats.filter(lambda x: x[1][1] >= min_ratings)

# Find the highest-rated movies

3
highest_rated_movies = qualified_movies.sortBy(lambda x: x[1][0],
ascending=False).take(10)

print(f"Highest-rated movies with at least {min_ratings} ratings: {highest_rated_movies}")

You might also like