Who sold more albums? Kanye or Drake?

image

Summary of the Project

Developed a Tableau visualization that showcased my ability to accomplish the following things:

  • Build a dashboard only using 1 floating object
  • Utilized my creativity in design only using Tableau
  • Find a unique data source that no one has done before
  • Make it about something I really cared about

So with those things in mind, I developed a dashboard that showcased Kanye vs Drake's first week sales.

Problem Statement

Kanye Vs Drake has been a hot debated topic amongst the internet (see photo below). Everyone wanted to know this one question - Who had the better album?

image

Solution

Process

Tools Utilized:

  • Tableau
  • Excel
  • Spotify

Dataset:

  • Spotify Charts
  • Wikipedia
  1. Brainstorming - Think about some ideas and key questions

I started this project before I learned Figma, so I didn't know how to build a dashboard using graphics at this point. Instead of sketching something out like I usually do, I brainstormed for 4 hours on metrics for success:

Brainstorming Questions:

How do you determine which album is better?

  • Reviews.
  • Album Sales / Streams.

I thought that it would difficult to measure reviews for the following reasons:

  • It would be difficult to aggregate a list of reviews
  • All reviews are not the same
    • Since all reviews are not the same, how do you weigh the significance of relevant reviews?

Due to these reasons, I wanted to focus on Album Sales / Streams. The next question would be how would you find this information?

How would you find Album / Sales / Streams?

  • Youtube
  • Spotify
  • Apple Music
  • Radio
  • Billboard Charts

The easiest way to find this information is through Spotify Charts. Other datasources were too costly to utilize.

Spotify Charts was the best data source for the following resources:

  • Free
  • Relevant Data
  1. Brainstorming - Create the template dashboard using Tableau

I like to design things with the end in mind.

This was my main inspiration for the dashboard. Once I had this in mind, I knew what I wanted to plan for the visualization

image
  1. Executing - Clean the data / Transform the Data / Data Integrity using Excel

Applications Used:

Excel

Tableau

Spotify Charts

Wikipedia

Excel Functions Used:

Index Match

Downloading the data

The first thing that I needed was relevant data. In order to compare the albums, I needed to find the start date for both artists.

Donda started on August 29th, 2021

CLB released on September 3rd, 2021

I downloaded the Global Daily Top 200 tracks based on those start dates.

image

Cleaning the data

I knew that I had to and clean and transform the data based on the CSV file. There was a lot of irrelevant information that was in this CSV. Since I worked with data visualizations before, I had an idea of what would be relevant and important information.

I thought it useful to have the following columns:

  • Track Name
  • Track Name 2
  • Num_Plays
  • Date Played
  • Time Length
  • URL (Spotify URL)

I knew that I would be combining all of the data into one spreadsheet and I needed a way to pull the following categories:

  • Num_Plays
  • URL

In order to get this data, I would have to Index Match Artist and Track Name. I ended up adding a Time Length Column, but I didn't really end up using it. I pulled that data from Wikipedia.

I noticed that some Track Names were messed up when I was cleaning the data. Some of the track names looked like this "Yebba’s Heartbreak" instead of 'Yebbas Heartbreak". In order utilize Index Match, I would still have to use the Spotify listed Track Name to pull the data.

In order to make sure that my information was clean, I decided to create a second Track Name column called Track Name 2 which housed the correct track name.

I removed the following in the Track Name 2:

  • apostrophes
  • brackets
  • Features
  • with
  • and
  • other misc.

This would save me time in Tableau, so I didn't have to change any alias.

Spotify Charts CSV

image

Time Length Per Track

image

Transforming the data

I knew how the data should look like in order to visualize it.

I needed to accomplish the following tasks to transform the data.

  1. Copy columns A2:D49 12x
    1. 8/29/2021 (Kanye's first album dropped) > 9/9/2021 (7 Days after Drake's album dropped) = 12 days
  2. Add the corresponding dat for Column E.

In order to make sure that I inputted the correct data for Column E, I just used the following excel function.

Count If

This function checks how many times I've inputted the criteria.

After I created the visualization, I noticed that some tableau visualzations had spotify players. I wanted to add a spotify player of my own, so I had to make some changes to the URL It looks like it would only show up with a player if I added an embed in the link.

URL from Spotify

Edited URL

image
image

Data Integrity

There are Null Values in this project for the following reasons:

  • Ok Ok and Jail Pt 2 were not listed on Spotify Charts during Kanye's release date
  • If the track fell off the top 200, it would not show up on Spotify Charts.

  1. Executing - Analyze the data using Tableau

Thought Process:

I really wanted to showcase my knowledge of using spark lines, bar graphs, and calculations in this project. I thought it would be cool to show the Top 5 Tracks of each artist, the total number of streams for each album, and the daily streams of each album. This dashboard would only be tracking the first 7 days of each project.

Here is a list of interesting insights that could be unlocked with BI tools:

  • What are the top tracks from each artist?
  • Can I listen to the top tracks from each artist?
  • How did each artist preform on a day to day basis?
  • How much of the Spotify Streams translated into Album sales?

Here is a list of filters / parameters / categories / insights that I wanted to include in the project:

  • Total Streams of Donda
  • Daily Streams of Donda
  • Total Streams of CLB
  • Total Streams of Donda
  • Album Sales calculation
  • Track per Artist filter

Once I had a list of what I wanted to accomplish, the rest of the project came down to execution and putting it in the right places.

Further Exploration

Here are two questions that I wanted to explore further:

  1. Which album had better retention rate?

This question would measure how much people still listened to the album after it's release date.

  1. What percent of Spotify streams make up the total number of streams across all streaming platforms?

This question would measure how the power of Spotify in comparison to the other streaming giants (Youtube, Tidal, Apple Music)

Feedback from others

  1. Visualize how the derivative of the views (ie, dropoff by day)
  2. Visualize the sales:

    3. What effects streaming

    Powered by ads

    Premium vs Free