The Olivia Rodrigo Project


Network Visualization Analysis


Noun Visualization

For our Noun Visualization, we took the top 30 nouns throughout all of Olivia's, Taylor's, Billie's, and Sabrina's song and created a Network Visualization based off the data. Each line connects to a word that is used by an artist. The thickness of line depends on how much the artist uses the word. The most used noun by these artist was "time" as it was used 191 times. Individually it was used by Billie "24" times, Olivia "25" times, Sabrina "86" times, and Taylor "51" times. When looking at the line connections, one can see the thickness variation according to their use by each artist. The use of nouns like "time," "night," "day," and "daylight" indicates a focus on time of day. The prominence of "love," "baby," "friend," and "boy" suggests a recurring theme of relationships and love in their music. Specifically for Olivia, she used the word "love" "59" times throughout her songs. This is interesting as she had the least amount of songs compared to Taylor, Sabrina, and Billie. Their "love" count was "99," "56," and "44" respectively. Olivia seems to put heavy emphasis on love throughout her songs.

Problems: If you look closely in the Network Visualization, some of the output is incorrect. Through using SpaCy and NLP(Natural Language Processing) we gathered the NOUN data, however, some of the NOUNs were not nouns. In our data visualization, it includes "da" and "-" which are not nouns.




Adjective Visualization

This Adjective Visualization showcases the consistent use of the top 30 adjectives across all artists. Heavily used adjectives include words like "bad," "good," and "well." These words specifically highlight the tendency of these artists to describe something as "bad," "good," or "well." Olivia uses the word "good" "27" times, while Billie, Sabrina, and Taylor use it "18," "35," and "19." This is a high count considering Olivia has 26 songs compared to Taylor: 46, Sabrina: 30, and Billie: 37.


network adjective

Problems: We struggled to get this visualization to work with PyVis. The closeness centrality is also working incorrectly, however, we are looking to fix this by our project completion date. Moreover, we had the same problem as we did in our Noun Visualization. There was a tendency for SpaCy and NLP to pick up words like "la." As far as the accuracy of the Adjective Visualization, we cannot be certain. Some words like "okay" and "only" could have been picked up in a different context that was not using it as an adjective.

Antconc Keyness (Effect & Likelihood) Visualization Analysis


Keyness (Effect)

Using AntConc's keyword analysis feature, we were able to find  statistically significant keywords in Olivia Rodrigo's song lyrics.  After identifying these top words, we looked at their Keyness (Effect),  which indicates how strong each of their associations are in Olivia's lyrics compared to Taylor's, Billie's, and Sabrina's combined lyrics. To visualize the stronger-associated words, the word cloud makes them appear larger than the weaker ones. The word "I" had the highest Keyness (Effect) at 0.127, followed by "and" (0.046), "so" (0.021), "all" (0.020), "ah" (0.014), "him" (0.012), "back" (0.010), "her" (0.08), and "she" (0.08). Words like "I", "him", "her", and "she", suggest that Olivia's lyrics focus primarily on personal experiences and relationships. "so", "and", "all", and "back" are words that indicate the storytelling side of her lyrics, they help to convey a certain mood or atmosphere.

keyness effect


Keyness (Likelihood)

The “Keyness” (Likelihood) results indicate the likelihood of each word being a key term in Olivia Rodrigo’s lyrics compared to the combined lyrics of the other artists in her genre. The word cloud visualizes this by not only making the higher values bigger, but also by its color gradient that turns from dark blue to pink as the likelihood values get lower. The word with the highest Keyness (Likelihood) value was “him” (94.186), followed by “I” (71.399), “her” (67.193), and “logical” (48.680). While the findings from the Keyness (Effect) can ultimately be inferred just by listening to Olivia's songs, the Keyness (Likelihood) results truly set apart Olivia's songwriting style from other pop female artists. The exceptionally high likelihood values of the words "him", "I", "her", and "logical" indicate just how prevelant and relied on they are by Olivia compared to the other artists. These words often correlate with themes of love and heartbreak, suggesting that these specific themes are more central for Olivia compared to Billie, Taylor, and Sabrina.

keyness likelihood

Additional Antconc Finding

Although the following information is not a visualization, we've decided to make note of it here as it provides additional insight into our findings. After analyzing our output data that was visualized in our PyVis noun network, Olivia's frequent use of the word "love," despite her lower song count compared to the other artists, was intriguing. We decided to take this information to AntConc, looking specifically at the "NormRange" for the word "love" with the goal of finding how statistically significant it is relative to each artist's song collection. Even with Olivia having the least amount of collected songs amongst the other artists, her "NormRange" for the word "love" was the highest at 0.630, followed by Billie at 0.459, Taylor at 0.370, and Sabrina at 0.233. This information further supports our previous reasonings that Olivia's themes are greatly influenced by relationships, likely exploring deeper emotions.

SVG Graph


This graph illustrates the song count per album for each artist. By examining the number of songs in each album, we gain insights into the volume of content contributed to this project and the typical output of each artist. Given that Olivia only has two complete albums along with an extension of her album 'GUTS', our ability to process her songs was constrained.

Problems:Completing the SVG graph was challenging in some ways as our code was reading incorrect numbers. It was producing the wrong count of songs per each album. To fix this we opened the SVG to input the correct numbers.

Back to Top

Created for the DIGIT 210: Text-analysis class at Penn State Behrend

Creative Commons License