TidyTuesday dataset of 2025-08-26

import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns

billboard = pd.read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/main/data/2025/2025-08-26/billboard.csv')
topics = pd.read_csv('https://raw.githubusercontent.com/rfordatascience/tidytuesday/main/data/2025/2025-08-26/topics.csv')

billboard

billboard.columns.values

array(['song', 'artist', 'date', 'weeks_at_number_one', 'non_consecutive',
'rating_1', 'rating_2', 'rating_3', 'overall_rating',
'divisiveness', 'label', 'parent_label', 'cdr_genre', 'cdr_style',
'discogs_genre', 'discogs_style', 'artist_structure',
'featured_artists', 'multiple_lead_vocalists',
'group_named_after_non_lead_singer', 'talent_contestant',
'posthumous', 'artist_place_of_origin', 'front_person_age',
'artist_male', 'artist_white', 'artist_black', 'songwriters',
'songwriters_w_o_interpolation_sample_credits', 'songwriter_male',
'songwriter_white', 'artist_is_a_songwriter',
'artist_is_only_songwriter', 'producers', 'producer_male',
'producer_white', 'artist_is_a_producer',
'artist_is_only_producer', 'songwriter_is_a_producer',
'time_signature', 'keys', 'simplified_key', 'bpm', 'energy',
'danceability', 'happiness', 'loudness_d_b', 'acousticness',
'vocally_based', 'bass_based', 'guitar_based',
'piano_keyboard_based', 'orchestral_strings', 'horns_winds',
'accordion', 'banjo', 'bongos', 'clarinet', 'cowbell',
'falsetto_vocal', 'flute_piccolo', 'handclaps_snaps', 'harmonica',
'human_whistling', 'kazoo', 'mandolin', 'pedal_lap_steel',
'ocarina', 'saxophone', 'sitar', 'trumpet', 'ukulele', 'violin',
'sound_effects', 'song_structure', 'rap_verse_in_a_non_rap_song',
'length_sec', 'instrumental', 'instrumental_length_sec',
'intro_length_sec', 'vocal_introduction',
'free_time_vocal_introduction', 'fade_out', 'live', 'cover',
'sample', 'interpolation', 'inspired_by_a_different_song',
'lyrics', 'lyrical_topic', 'lyrical_narrative', 'spoken_word',
'explicit', 'foreign_language', 'written_for_a_play',
'featured_in_a_then_contemporary_play', 'written_for_a_film',
'featured_in_a_then_contemporary_film', 'written_for_a_t_v_show',
'featured_in_a_then_contemporary_t_v_show',
'associated_with_dance', 'topped_the_charts_by_multiple_artist',
'double_a_side', 'eurovision_entry', 'u_s_artwork'], dtype=object)

billboard['date'] = pd.to_datetime(billboard['date'])
# add a new  col date2 having three categories Before 75, 75-2000, After 2000
billboard['date2'] = pd.cut(billboard['date'].dt.year, bins=[0, 1975, 2000, 2025], labels=['Before 1975', '1975-2000', 'After 2000'])
billboard['date2'].value_counts()

date2
1975-2000 484
Before 1975 383
After 2000 310
Name: count, dtype: int64

billboard['simplified_key_mod'] = billboard['simplified_key'].replace('Multiple Keys', 'multi')
sk_values = billboard['simplified_key_mod'].value_counts()
sk_values

simplified_key_mod
C 120
multi 96
G 83
F 76
D 70
E 69
A 67
Db 57
Ab 49
Bb 48
Eb 42
Am 40
Fm 34
Cm 32
Gm 31
Dbm 29
Dm 29
Gb 29
B 29
Bbm 28
Em 27
Ebm 26
Bm 25
Gbm 21
Abm 20
Name: count, dtype: int64

fig,ax = plt.subplots(figsize=(10,5))

sns.swarmplot(data=billboard, x='date2', y='simplified_key_mod', hue='weeks_at_number_one', \
ax=ax, alpha=0.75)
ax.legend(loc='lower center', ncol=6, bbox_to_anchor=(0.5, -0.2), \
frameon=False, title="")
ax.text(-0.55,29.4, "Weeks at Number One →")
sns.despine(bottom=True)
ax.tick_params(axis='x', which='both', length=0)

plt.xlabel('')
plt.ylabel('Key (simplified)')
plt.title("Changing key preference over the years in Billboard hits.", fontfamily='serif', \
         fontsize=14)
plt.savefig('billboard_hits.png', dpi=300, bbox_inches='tight')
plt.show()

sns.pairplot(billboard[['bpm', 'energy','danceability', 'happiness','weeks_at_number_one']], hue='weeks_at_number_one',\
corner=True)
plt.show()

Update

A revised visualization for this data was submitted to Plotnine contest 2025 (Repo).