<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Chantell Akuffo on Medium]]></title>
        <description><![CDATA[Stories by Chantell Akuffo on Medium]]></description>
        <link>https://medium.com/@chantellakuffo?source=rss-a8279231564------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/2*fRPBprxsZsUBCAykLX-m1A.jpeg</url>
            <title>Stories by Chantell Akuffo on Medium</title>
            <link>https://medium.com/@chantellakuffo?source=rss-a8279231564------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 05:36:42 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@chantellakuffo/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[What Makes a Song Popular]]></title>
            <link>https://medium.com/@chantellakuffo/what-makes-a-song-popular-cae3c6ff66c2?source=rss-a8279231564------2</link>
            <guid isPermaLink="false">https://medium.com/p/cae3c6ff66c2</guid>
            <category><![CDATA[spotify]]></category>
            <category><![CDATA[music]]></category>
            <category><![CDATA[big-data]]></category>
            <category><![CDATA[data-visualization]]></category>
            <category><![CDATA[billboard]]></category>
            <dc:creator><![CDATA[Chantell Akuffo]]></dc:creator>
            <pubDate>Mon, 02 Dec 2019 20:25:05 GMT</pubDate>
            <atom:updated>2019-12-03T12:31:34.561Z</atom:updated>
            <content:encoded><![CDATA[<p>by: Chantell Akuffo, Liya Wu and Alison Jacobs</p><p>Have you ever wondered what makes a song popular? How do you determine which songs will make it onto the Top 100 charts? Is there a way to construct the audio features so that you ensure you have a “hit song”? Through this post, we will analyze Billboard’s Hot 100 List every week for the past five years and retrieve its audio features through Spotify.</p><p>By using the Billboard.py API, scraping the data from Billboard was as simple as looping through our date frame (January 2014 to November 2019) every week and indexing into the chart to create a set of song titles and artists.</p><p>After retrieving our data set( the weekly top 100 songs and the corresponding artists from Billboard.py) and analysing it, we decided to go a step further and generate each song’s URI as well as each artist’s URI using spotipy. With this, we were able to find the audio features such as danceability, loudness, energy, instrumentalness, and acousticness of each song.</p><p>With the songs’ URIs, we were able to access the genres of each song and find the top 5 most occuring genre. The top 5 genres for our dataset were: pop, teen pop, dance pop, contemporary country, and country.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/711/1*Po-kfYX7CIPyi1kHB6EsWQ.png" /></figure><p>We further narrowed down our song list by only looking at the songs that were in the top 5 genre categories. Then by accessing the Spotify Api, we were able to find statistics about each specific song such as the danceability, energy tempo, instrumentalness, and acousticness. Based on the genre of each song, we added their audio features to a specific list which then allowed us to find the average value for each category. This enables us to predict the best value for each feature for a specific genre to make it on the Hot 100 list.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/361/1*gL2lee1ZlCtjz0dMkj12rQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/399/1*2SL_LzQrExkgr73yveDMCw.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/392/1*MKekemdR9X2r-rwwIdYPeg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/376/1*zd2de4FuP0_4Kpn_TY9rUQ.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/380/1*i80yKIHOlX0744kenpw6RQ.png" /></figure><p>We then took the value of each category for every genre and graphed the results to draw comparisons. The resulting graphs are displayed below.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/752/0*KAWZv9mOpyyXi0HI" /></figure><p>Danceability is a float value from 0.0 to 1.0 which represents how suitable a track is for dancing based on different musical elements such as tempo, rhythm stability, beat strength, and overall regularity with 0.0 being least danceable and 1.0 being most danceable.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/716/1*80pxU3AUnDJbN-Cy3y6IFg.png" /></figure><p>Energy is a perceptual measure of intensity and activity of a song as a float value from 0.0 to 1.0 with 0.0 being least energetic.These perceptual features include dynamic range, perceived loudness, timbre, onset rate, and general entropy. Typically, energetic songs feel fast, loud, and noisy. Hence, death metal will have a high energy whereas a Bach prelude will have a low energy score.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/715/1*OrJYXEvxSYm-NcKG4T1ZDg.png" /></figure><p>Loudness is basically a measure of the overall loudness of a track in decibels(dB). Loudness is the quality of a song that is the primary psychological correlation of physical strength(amplitude). The loudness of songs typically range between -60 and 0 dB.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/644/1*hus0f044qMttpH6seEYodQ.png" /></figure><p>Tempo is the overall estimated time signature of a song represented as an integer. The time signature is measured in meters and it is a notational convention to specify how many beats are in each bar.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/703/1*o4ze2XmJUTtMiaISjn_GlQ.png" /></figure><p>Instrumentalness is the prediction of whether a song contains no vocals. “Ooh” and “aah” sounds are treated as instrumentals. Hence, rap or spoken word songs are clearly vocal and as such will have an instrumentalness value of closer to 1.0. Values above 0.5 are intended to represent instrumental songs.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/689/1*GF5jJqkFBkhDXDCjET7y1g.png" /></figure><p>Acousticness measures whether a song is acoustic from a range of 0.0 to 1.0 with 1.0 representing a high confidence that the song is acoustic.</p><p>We further desired to gain more insight into data so a dictionary was created to organize the data and to map the artist to the number of unique songs they have on the top 100 since January 2015. Drake had an overwhelming number of songs on the top 100, some of the songs on the billboard are including Hotline Bling, God’s Plan and Nonstop. The graph below depicts the artist and the amount of unique songs they have on the Hot 100 list over the past 5 years.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/743/1*BWk1XS1ctLq9Dz_KDf-Nqg.png" /></figure><p>Furthermore, the graph below depicts which songs were the most popular, and appeared on the top 100 most times over the past 5 years. The top 20 songs are shown below with the top songs being Perfect by Ed Sheeran, and then Worth It by Fifth Harmony.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/718/1*1v1ClpOxZCecxXrer9728g.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/632/1*J1S7jf1Ddla1CG2jx--KXw.png" /></figure><p>Moving forward, a TF/IDF analysis was created using the words present in all 25,397 songs. However, due to the large amount of data, only the top 500 songs were used and analyzed by their lyrics. The graphs below display the top three songs with the most words. Those songs are Bad and boujee, Walk It Talk It and I’m the One. The graphs further display the top 10 words and their TF/IDF value.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/516/1*2o3uIU3PUxDYjym43Lvnvg.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/560/1*NzenNo-dHRBIhlsPPZCbmA.png" /></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/589/1*9Drub5WZDWbxn6NgwOIaWw.png" /></figure><p>By using the Billboard and Spotify APIs, we were able to successfully predict what it takes for a song to be on the Hot 100 Chart. The most successful songs were in the categories of Pop, Dance Pop, Teen Pop, Contemporary Country, and Country. Through analyzing each song in their respective genre, we found the average values for each audio feature: danceability, energy, tempo, instrumentalness, and acousticness. These values could be used to help create a song that has a higher chance of being featured on the Hot 100 List. As seen above, Drake was the artist with the most songs on the top 100 over the past five years, and Perfect by Ed Sheeran was the most popular song. Also, after doing a TF/IDF analysis, the graphs above show the most signficant words for the three songs with the most words in the data set. This is significant because it highlights which words are significant in a song in relation to all the words in the dataset.</p><p>Link to all our <a href="https://github.com/chantella/spotify-analysis">Github</a> repository containing all the code we used in this project.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=cae3c6ff66c2" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>