A startup called Sparkify wants to analyze the data they've been collecting on songs and user activity on their new music streaming app. The analytics team is particularly interested in understanding what songs users are listening to. Currently, they don't have an easy way to query their data, which resides in a directory of JSON logs on user activity on the app, as well as a directory with JSON metadata on the songs in their app.
The aim of this project is to create a Postgres database with tables designed to optimize queries on song play analysis.
- Your role is to create a database schema and ETL pipeline for this analysis. You'll be able to test your database and ETL pipeline by running queries given to you by the analytics team from Sparkify and compare your results with their expected results.
In this project, you'll apply what you've learned on data modeling with Postgres and build an ETL pipeline using Python. To complete the project, you will need to define fact and dimension tables for a star schema for a particular analytic focus, and write an ETL pipeline that transfers data from files in two local directories into these tables in Postgres using Python and SQL.
- The Database schema is a star schema optimized for queries on song play analysis and this includes:
- Fact Table which consists of this table:
songplays - records in log data associated with song plays i.e. records with page NextSong
consisting of following columns:songplay_id, start_time, user_id, level, song_id, artist_id, session_id, location, user_agent
- Dimension Tables which consists of the following tables:
users - users in the app
consisting of following columns:user_id, first_name, last_name, gender, level
songs - songs in music database
consisting of following columns:song_id, title, artist_id, year, duration
artists - artists in music database
consisting of following columns:artist_id, name, location, latitude, longitude
time - timestamps of records in songplays broken down into specific units
consisting of following columns:start_time, hour, day, week, month, year, weekday
- Song Dataset
The first dataset is a subset of real data from the Million Song Dataset. Each file is in JSON format and contains metadata about a song and the artist of that song. The files are partitioned by the first three letters of each song's track ID. For example, here are filepaths to two files in this dataset: song_data/A/B/C/TRABCEI128F424C983.json; song_data/A/A/B/TRAABJL12903CDCF1A.json.
And below is an example of what a single song file, TRAABJL12903CDCF1A.json, looks like {"num_songs": 1, "artist_id": "ARJIE2Y1187B994AB7", "artist_latitude": null, "artist_longitude": null, "artist_location": "", "artist_name": "Line Renaud", "song_id": "SOUPIRU12A6D4FA1E1", "title": "Der Kleine Dompfaff", "duration": 152.92036, "year": 0}
- Log Dataset
The second dataset consists of log files in JSON format generated by this event simulator based on the songs in the dataset above. These simulate activity logs from a music streaming app based on specified configurations.
The log files in the dataset you'll be working with are partitioned by year and month. For example, here are filepaths to two files in this dataset: log_data/2018/11/2018-11-12-events.json, log_data/2018/11/2018-11-13-events.json.
-
First,
Drop Tables If they initially exist
by writingDROP Table statement in sql_queries.py module
written for specific tables. -
After dropping existing tables,
Create Tables
by writingCREATE Table statement in sql_queries.py module
written for specific tables. -
Thereafter, Insert the required columns/records in the created tables, writing
Insert statement for each table created in sql_queries.py module.
-
Create_table.py is ran to create the database, drop and create the tables
. -
Furthermore,
test.ipynb
is ran to ascertain the tables were created. -
Building ETL pipeline (extract, transform and load) process consist of:
-
First, Extracting the song files from the song_data directory, since the song_files are in json format, we add .json at the end of each file
-
Read the json song file into a pandas Dataframe.
-
Extract the required columns for the song and artist table, and further load it into these tables
-
Second, Extract the log files from the log_data directory, since the log_files are in json format, we add .json at the end of each file
-
Read the json log file into a pandas Dataframe
-
Transform the time column into a datetime format and then further extract various features for it such as hour, week, month, year, weekday and start time.
-
Extract the required column from the log dataframe including the transformed time columns, and further load it into the user and time tables.
-
Song id and artist id selected from the song and artist tables created earlier, together with some columns from the log dataframe are loaded into the songplay table.
-
After completing the build-up of the ETL pipeline, the code are transferred and implemented in
etl.py
-
After the implementation, run the
test.ipynb
to ascertain your records were inserted.