The purpose of this project is to compare the music preferences in the cities of Springfield and Shelbyville. We will use data provided by Yandex.music to test hypotheses on user behavior.
Open the data on Yandex.Music and explore it.
# importing pandas
import pandas as pd
# read dataset and store to df
df = pd.read_csv('datasets/music_project_en.csv')
# ensure proper reading of file
df.head()
| userID | Track | artist | genre | City | time | Day | |
|---|---|---|---|---|---|---|---|
| 0 | FFB692EC | Kamigata To Boots | The Mass Missile | rock | Shelbyville | 20:28:33 | Wednesday |
| 1 | 55204538 | Delayed Because of Accident | Andreas Rönnberg | rock | Springfield | 14:07:09 | Friday |
| 2 | 20EC38 | Funiculì funiculà | Mario Lanza | pop | Shelbyville | 20:58:07 | Wednesday |
| 3 | A3DD03C9 | Dragons in the Sunset | Fire + Ice | folk | Shelbyville | 08:37:09 | Monday |
| 4 | E2DC1FAE | Soul People | Space Echo | dance | Springfield | 08:34:34 | Monday |
# description of the data
df.describe()
| userID | Track | artist | genre | City | time | Day | |
|---|---|---|---|---|---|---|---|
| count | 65079 | 63736 | 57512 | 63881 | 65079 | 65079 | 65079 |
| unique | 41748 | 39666 | 37806 | 268 | 2 | 20392 | 3 |
| top | A8AE9169 | Brand | Kartvelli | pop | Springfield | 08:14:07 | Friday |
| freq | 76 | 136 | 136 | 8850 | 45360 | 14 | 23149 |
# obtaining the first 10 rows from the df table
df.head(10)
| userID | Track | artist | genre | City | time | Day | |
|---|---|---|---|---|---|---|---|
| 0 | FFB692EC | Kamigata To Boots | The Mass Missile | rock | Shelbyville | 20:28:33 | Wednesday |
| 1 | 55204538 | Delayed Because of Accident | Andreas Rönnberg | rock | Springfield | 14:07:09 | Friday |
| 2 | 20EC38 | Funiculì funiculà | Mario Lanza | pop | Shelbyville | 20:58:07 | Wednesday |
| 3 | A3DD03C9 | Dragons in the Sunset | Fire + Ice | folk | Shelbyville | 08:37:09 | Monday |
| 4 | E2DC1FAE | Soul People | Space Echo | dance | Springfield | 08:34:34 | Monday |
| 5 | 842029A1 | Chains | Obladaet | rusrap | Shelbyville | 13:09:41 | Friday |
| 6 | 4CB90AA5 | True | Roman Messer | dance | Springfield | 13:00:07 | Wednesday |
| 7 | F03E1C1F | Feeling This Way | Polina Griffith | dance | Springfield | 20:47:49 | Wednesday |
| 8 | 8FA1D3BE | L’estate | Julia Dalia | ruspop | Springfield | 09:17:40 | Friday |
| 9 | E772D5C0 | Pessimist | NaN | dance | Shelbyville | 21:20:49 | Wednesday |
# obtaining general information about the data in df
df.info()
<class 'pandas.core.frame.DataFrame'> RangeIndex: 65079 entries, 0 to 65078 Data columns (total 7 columns): # Column Non-Null Count Dtype --- ------ -------------- ----- 0 userID 65079 non-null object 1 Track 63736 non-null object 2 artist 57512 non-null object 3 genre 63881 non-null object 4 City 65079 non-null object 5 time 65079 non-null object 6 Day 65079 non-null object dtypes: object(7) memory usage: 3.5+ MB
Correct the formatting in the column headers and deal with the missing values. Then, check whether there are duplicates in the data.
# the list of column names in the df table
df.columns
Index([' userID', 'Track', 'artist', 'genre', ' City ', 'time', 'Day'], dtype='object')
# renaming columns
df = df.rename(columns={' userID' : 'user_id', 'Track' : 'track', ' City ' : 'city', 'Day' : 'day'})
# checking result: the list of column names
df.columns
Index(['user_id', 'track', 'artist', 'genre', 'city', 'time', 'day'], dtype='object')
# calculating missing values
df.isna().sum()
user_id 0 track 1343 artist 7567 genre 1198 city 0 time 0 day 0 dtype: int64
# looping over column names and replacing missing values with 'unknown'
columns_to_replace = df.columns
for column in columns_to_replace :
df[column].fillna('unknown', inplace=True)
#checking to see if NaN still exists in data set
df.head(16)
| user_id | track | artist | genre | city | time | day | |
|---|---|---|---|---|---|---|---|
| 0 | FFB692EC | Kamigata To Boots | The Mass Missile | rock | Shelbyville | 20:28:33 | Wednesday |
| 1 | 55204538 | Delayed Because of Accident | Andreas Rönnberg | rock | Springfield | 14:07:09 | Friday |
| 2 | 20EC38 | Funiculì funiculà | Mario Lanza | pop | Shelbyville | 20:58:07 | Wednesday |
| 3 | A3DD03C9 | Dragons in the Sunset | Fire + Ice | folk | Shelbyville | 08:37:09 | Monday |
| 4 | E2DC1FAE | Soul People | Space Echo | dance | Springfield | 08:34:34 | Monday |
| 5 | 842029A1 | Chains | Obladaet | rusrap | Shelbyville | 13:09:41 | Friday |
| 6 | 4CB90AA5 | True | Roman Messer | dance | Springfield | 13:00:07 | Wednesday |
| 7 | F03E1C1F | Feeling This Way | Polina Griffith | dance | Springfield | 20:47:49 | Wednesday |
| 8 | 8FA1D3BE | L’estate | Julia Dalia | ruspop | Springfield | 09:17:40 | Friday |
| 9 | E772D5C0 | Pessimist | unknown | dance | Shelbyville | 21:20:49 | Wednesday |
| 10 | BC5A3A29 | Gool la Mita | Shireen Abdul Wahab | world | Springfield | 14:08:42 | Monday |
| 11 | 8B5192C0 | Is There Anybody Out There? (Panoramic Paralys... | Pink Floyd Floydhead | electronic | Springfield | 13:47:49 | Monday |
| 12 | FF3FD2BD | Truth | Bamboo | pop | Springfield | 09:19:49 | Monday |
| 13 | CC782B0F | After School Special | Detroit Grand Pubahs | dance | Springfield | 20:04:12 | Friday |
| 14 | 94EB25C2 | Make Love Whenever You Can | Arabesque | pop | Shelbyville | 13:22:08 | Wednesday |
| 15 | E3C5756F | unknown | unknown | unknown | Springfield | 09:24:51 | Monday |
# counting missing values
df.isna().sum()
user_id 0 track 0 artist 0 genre 0 city 0 time 0 day 0 dtype: int64
Find the number of obvious duplicates in the table
# counting clear duplicates
print(df.duplicated().sum())
3826
# removing obvious duplicates
df= df.drop_duplicates()
# checking for duplicates
df.duplicated().sum()
0
# looking at tail of data
df.tail()
| user_id | track | artist | genre | city | time | day | |
|---|---|---|---|---|---|---|---|
| 65074 | 729CBB09 | My Name | McLean | rnb | Springfield | 13:32:28 | Wednesday |
| 65075 | D08D4A55 | Maybe One Day (feat. Black Spade) | Blu & Exile | hip | Shelbyville | 10:00:00 | Monday |
| 65076 | C5E3A0D5 | Jalopiina | unknown | industrial | Springfield | 20:09:26 | Friday |
| 65077 | 321D0506 | Freight Train | Chas McDevitt | rock | Springfield | 21:43:59 | Friday |
| 65078 | 3A64EF84 | Tell Me Sweet Little Lies | Monica Lopez | country | Springfield | 21:59:46 | Friday |
# resetting index after dropping rows, look at tail
df.reset_index(drop=True).tail()
| user_id | track | artist | genre | city | time | day | |
|---|---|---|---|---|---|---|---|
| 61248 | 729CBB09 | My Name | McLean | rnb | Springfield | 13:32:28 | Wednesday |
| 61249 | D08D4A55 | Maybe One Day (feat. Black Spade) | Blu & Exile | hip | Shelbyville | 10:00:00 | Monday |
| 61250 | C5E3A0D5 | Jalopiina | unknown | industrial | Springfield | 20:09:26 | Friday |
| 61251 | 321D0506 | Freight Train | Chas McDevitt | rock | Springfield | 21:43:59 | Friday |
| 61252 | 3A64EF84 | Tell Me Sweet Little Lies | Monica Lopez | country | Springfield | 21:59:46 | Friday |
# viewing unique genre names
sorted_df = df.sort_values(by='genre', ascending=True)
print(sorted_df['genre'].unique())
['acid' 'acoustic' 'action' 'adult' 'africa' 'afrikaans' 'alternative' 'ambient' 'americana' 'animated' 'anime' 'arabesk' 'arabic' 'arena' 'argentinetango' 'art' 'audiobook' 'avantgarde' 'axé' 'baile' 'balkan' 'beats' 'bigroom' 'black' 'bluegrass' 'blues' 'bollywood' 'bossa' 'brazilian' 'breakbeat' 'breaks' 'broadway' 'cantautori' 'cantopop' 'canzone' 'caribbean' 'caucasian' 'celtic' 'chamber' 'children' 'chill' 'chinese' 'choral' 'christian' 'christmas' 'classical' 'classicmetal' 'club' 'colombian' 'comedy' 'conjazz' 'contemporary' 'country' 'cuban' 'dance' 'dancehall' 'dancepop' 'dark' 'death' 'deep' 'deutschrock' 'deutschspr' 'dirty' 'disco' 'dnb' 'documentary' 'downbeat' 'downtempo' 'drum' 'dub' 'dubstep' 'eastern' 'easy' 'electronic' 'electropop' 'emo' 'entehno' 'epicmetal' 'estrada' 'ethnic' 'eurofolk' 'european' 'experimental' 'extrememetal' 'fado' 'film' 'fitness' 'flamenco' 'folk' 'folklore' 'folkmetal' 'folkrock' 'folktronica' 'forró' 'frankreich' 'französisch' 'french' 'funk' 'future' 'gangsta' 'garage' 'german' 'ghazal' 'gitarre' 'glitch' 'gospel' 'gothic' 'grime' 'grunge' 'gypsy' 'handsup' "hard'n'heavy" 'hardcore' 'hardstyle' 'hardtechno' 'hip' 'hip-hop' 'hiphop' 'historisch' 'holiday' 'hop' 'horror' 'house' 'idm' 'independent' 'indian' 'indie' 'indipop' 'industrial' 'inspirational' 'instrumental' 'international' 'irish' 'jam' 'japanese' 'jazz' 'jewish' 'jpop' 'jungle' 'k-pop' 'karadeniz' 'karaoke' 'kayokyoku' 'korean' 'laiko' 'latin' 'latino' 'leftfield' 'local' 'lounge' 'loungeelectronic' 'lovers' 'malaysian' 'mandopop' 'marschmusik' 'meditative' 'mediterranean' 'melodic' 'metal' 'metalcore' 'mexican' 'middle' 'minimal' 'miscellaneous' 'modern' 'mood' 'mpb' 'muslim' 'native' 'neoklassik' 'neue' 'new' 'newage' 'newwave' 'nu' 'nujazz' 'numetal' 'oceania' 'old' 'opera' 'orchestral' 'other' 'piano' 'pop' 'popelectronic' 'popeurodance' 'post' 'posthardcore' 'postrock' 'power' 'progmetal' 'progressive' 'psychedelic' 'punjabi' 'punk' 'quebecois' 'ragga' 'ram' 'rancheras' 'rap' 'rave' 'reggae' 'reggaeton' 'regional' 'relax' 'religious' 'retro' 'rhythm' 'rnb' 'rnr' 'rock' 'rockabilly' 'romance' 'roots' 'ruspop' 'rusrap' 'rusrock' 'salsa' 'samba' 'schlager' 'self' 'sertanejo' 'shoegazing' 'showtunes' 'singer' 'ska' 'slow' 'smooth' 'soul' 'soulful' 'sound' 'soundtrack' 'southern' 'specialty' 'speech' 'spiritual' 'sport' 'stonerrock' 'surf' 'swing' 'synthpop' 'sängerportrait' 'tango' 'tanzorchester' 'taraftar' 'tech' 'techno' 'thrash' 'top' 'traditional' 'tradjazz' 'trance' 'tribal' 'trip' 'triphop' 'tropical' 'türk' 'türkçe' 'unknown' 'urban' 'uzbek' 'variété' 'vi' 'videogame' 'vocal' 'western' 'world' 'worldbeat' 'ïîï']
# function for replacing implicit duplicates
def replace_wrong_genres(wrong_genres, correct_genres) :
for wrong_genre in wrong_genres :
df['genre'] =df['genre'].replace(wrong_genre, correct_genres)
# removing implicit duplicates
duplicates = [['hip'], ['hop'], ['hip-hop']]
new = 'hiphop'
replace_wrong_genres(duplicates, new)
print(df)
user_id track artist \
0 FFB692EC Kamigata To Boots The Mass Missile
1 55204538 Delayed Because of Accident Andreas Rönnberg
2 20EC38 Funiculì funiculà Mario Lanza
3 A3DD03C9 Dragons in the Sunset Fire + Ice
4 E2DC1FAE Soul People Space Echo
... ... ... ...
65074 729CBB09 My Name McLean
65075 D08D4A55 Maybe One Day (feat. Black Spade) Blu & Exile
65076 C5E3A0D5 Jalopiina unknown
65077 321D0506 Freight Train Chas McDevitt
65078 3A64EF84 Tell Me Sweet Little Lies Monica Lopez
genre city time day
0 rock Shelbyville 20:28:33 Wednesday
1 rock Springfield 14:07:09 Friday
2 pop Shelbyville 20:58:07 Wednesday
3 folk Shelbyville 08:37:09 Monday
4 dance Springfield 08:34:34 Monday
... ... ... ... ...
65074 rnb Springfield 13:32:28 Wednesday
65075 hiphop Shelbyville 10:00:00 Monday
65076 industrial Springfield 20:09:26 Friday
65077 rock Springfield 21:43:59 Friday
65078 country Springfield 21:59:46 Friday
[61253 rows x 7 columns]
Make sure the duplicate names were removed. Print the list of unique values from the 'genre' column:
# Checking for implicit duplicates
df = df.sort_values(by= 'genre', ascending=True) #sorting df by genre, alphabetical order,
print(df['genre'].unique()) #see hop changed to hiphop
['acid' 'acoustic' 'action' 'adult' 'africa' 'afrikaans' 'alternative' 'ambient' 'americana' 'animated' 'anime' 'arabesk' 'arabic' 'arena' 'argentinetango' 'art' 'audiobook' 'avantgarde' 'axé' 'baile' 'balkan' 'beats' 'bigroom' 'black' 'bluegrass' 'blues' 'bollywood' 'bossa' 'brazilian' 'breakbeat' 'breaks' 'broadway' 'cantautori' 'cantopop' 'canzone' 'caribbean' 'caucasian' 'celtic' 'chamber' 'children' 'chill' 'chinese' 'choral' 'christian' 'christmas' 'classical' 'classicmetal' 'club' 'colombian' 'comedy' 'conjazz' 'contemporary' 'country' 'cuban' 'dance' 'dancehall' 'dancepop' 'dark' 'death' 'deep' 'deutschrock' 'deutschspr' 'dirty' 'disco' 'dnb' 'documentary' 'downbeat' 'downtempo' 'drum' 'dub' 'dubstep' 'eastern' 'easy' 'electronic' 'electropop' 'emo' 'entehno' 'epicmetal' 'estrada' 'ethnic' 'eurofolk' 'european' 'experimental' 'extrememetal' 'fado' 'film' 'fitness' 'flamenco' 'folk' 'folklore' 'folkmetal' 'folkrock' 'folktronica' 'forró' 'frankreich' 'französisch' 'french' 'funk' 'future' 'gangsta' 'garage' 'german' 'ghazal' 'gitarre' 'glitch' 'gospel' 'gothic' 'grime' 'grunge' 'gypsy' 'handsup' "hard'n'heavy" 'hardcore' 'hardstyle' 'hardtechno' 'hiphop' 'historisch' 'holiday' 'horror' 'house' 'idm' 'independent' 'indian' 'indie' 'indipop' 'industrial' 'inspirational' 'instrumental' 'international' 'irish' 'jam' 'japanese' 'jazz' 'jewish' 'jpop' 'jungle' 'k-pop' 'karadeniz' 'karaoke' 'kayokyoku' 'korean' 'laiko' 'latin' 'latino' 'leftfield' 'local' 'lounge' 'loungeelectronic' 'lovers' 'malaysian' 'mandopop' 'marschmusik' 'meditative' 'mediterranean' 'melodic' 'metal' 'metalcore' 'mexican' 'middle' 'minimal' 'miscellaneous' 'modern' 'mood' 'mpb' 'muslim' 'native' 'neoklassik' 'neue' 'new' 'newage' 'newwave' 'nu' 'nujazz' 'numetal' 'oceania' 'old' 'opera' 'orchestral' 'other' 'piano' 'pop' 'popelectronic' 'popeurodance' 'post' 'posthardcore' 'postrock' 'power' 'progmetal' 'progressive' 'psychedelic' 'punjabi' 'punk' 'quebecois' 'ragga' 'ram' 'rancheras' 'rap' 'rave' 'reggae' 'reggaeton' 'regional' 'relax' 'religious' 'retro' 'rhythm' 'rnb' 'rnr' 'rock' 'rockabilly' 'romance' 'roots' 'ruspop' 'rusrap' 'rusrock' 'salsa' 'samba' 'schlager' 'self' 'sertanejo' 'shoegazing' 'showtunes' 'singer' 'ska' 'slow' 'smooth' 'soul' 'soulful' 'sound' 'soundtrack' 'southern' 'specialty' 'speech' 'spiritual' 'sport' 'stonerrock' 'surf' 'swing' 'synthpop' 'sängerportrait' 'tango' 'tanzorchester' 'taraftar' 'tech' 'techno' 'thrash' 'top' 'traditional' 'tradjazz' 'trance' 'tribal' 'trip' 'triphop' 'tropical' 'türk' 'türkçe' 'unknown' 'urban' 'uzbek' 'variété' 'vi' 'videogame' 'vocal' 'western' 'world' 'worldbeat' 'ïîï']
# checking for implicit duplicates
df.duplicated().sum()
0
#visual of df grouped by city, ensure correct values
print(df.groupby('city')['track'].count())
city Shelbyville 18512 Springfield 42741 Name: track, dtype: int64
# Counting up the tracks played in each city
tracks_by_city = df.groupby('city')['track'].count()
print(tracks_by_city)
city Shelbyville 18512 Springfield 42741 Name: track, dtype: int64
#visual of df grouped by day, ensure correct values
print(df.groupby('day')['track'].count())
day Friday 21840 Monday 21354 Wednesday 18059 Name: track, dtype: int64
# Calculating tracks played on each of the three days
tracks_by_day = df.groupby('day')['track'].count()
print(tracks_by_day)
day Friday 21840 Monday 21354 Wednesday 18059 Name: track, dtype: int64
#group df by day and city
day_by_city = df.groupby('day')['city'].count()
print(day_by_city)
day Friday 21840 Monday 21354 Wednesday 18059 Name: city, dtype: int64
#group df by city and day
city_by_day = df.groupby('city')['day'].count()
print(city_by_day)
city Shelbyville 18512 Springfield 42741 Name: day, dtype: int64
# number of values in day and city columns
df.loc[:,['day', 'city']].count()
day 61253 city 61253 dtype: int64
# filtering day and city columns
df_day_and_city = df.loc[:,['day', 'city']]
print(df_day_and_city)
day city 32883 Friday Springfield 3650 Monday Shelbyville 20644 Wednesday Springfield 39129 Monday Springfield 14811 Friday Shelbyville ... ... ... 6189 Friday Springfield 18268 Wednesday Springfield 28665 Monday Springfield 40493 Wednesday Shelbyville 8514 Friday Springfield [61253 rows x 2 columns]
There are substantially more song played in Springfield than Shelbyville; more than double. Among both cities, most songs are played on Friday, followed by Monday, and Then Wednesday.
# <creating the function number_tracks()>
def number_tracks(day, city) :
track_list = df.loc[(df["day"] == day) & (df["city"] == city), :]
track_list_count = len(track_list)
return track_list_count
city = 'city'
day = 'day'
column_values = number_tracks(day, city)
# the number of songs played in Springfield on Monday
print('Monday: Springfield')
number_tracks('Monday', 'Springfield')
Monday: Springfield
15740
# the number of songs played in Shelbyville on Monday
print('Monday: Shelbyville')
number_tracks('Monday', 'Shelbyville')
Monday: Shelbyville
5614
# the number of songs played in Springfield on Wednesday
print('Wednesday: Springfield')
number_tracks('Wednesday', 'Springfield')
Wednesday: Springfield
11056
# the number of songs played in Shelbyville on Wednesday
print('Wednesday: Shelbyville')
number_tracks('Wednesday', 'Shelbyville')
Wednesday: Shelbyville
7003
# the number of songs played in Springfield on Friday
print('Friday: Springfield')
number_tracks('Friday', 'Springfield')
Friday: Springfield
15945
# the number of songs played in Shelbyville on Friday
print('Friday: Shelbyville')
number_tracks('Friday', 'Shelbyville')
Friday: Shelbyville
5895
# table with results
import numpy as np
d = [['Springfield', 15740, 11056, 15945], ['Shelbyville', 5614, 7003, 5895]]
df2 = pd.DataFrame(np.array(d), columns=['city', 'monday', 'wednesday', 'friday'])
df2
| city | monday | wednesday | friday | |
|---|---|---|---|---|
| 0 | Springfield | 15740 | 11056 | 15945 |
| 1 | Shelbyville | 5614 | 7003 | 5895 |
As anticipated, the number of tracks played in Springfield is greater than that of Shelbyville, even when grouping by days. Friday and Monday are the days with most tracks played in Springfield. More tracks are played on Wednesday in Shelbyville, followed by Friday and Monday.
# create the spr_general table from the df rows,
# where the value in the 'city' column is 'Springfield'
spr_general = df[df['city']=='Springfield']
spr_general
| user_id | track | artist | genre | city | time | day | |
|---|---|---|---|---|---|---|---|
| 32883 | C40BC5DF | Shot in the Dark | Four80East | acid | Springfield | 09:44:12 | Friday |
| 20644 | B02ED3AB | Prelude – Ne Me Qui Te Pas | unknown | acoustic | Springfield | 13:31:03 | Wednesday |
| 39129 | 77D3A7D9 | Prelude – Ne Me Qui Te Pas | unknown | acoustic | Springfield | 14:27:34 | Monday |
| 60870 | 23C64044 | Don't Stop Me Now | Acoustic Heartstrings | acoustic | Springfield | 13:30:01 | Wednesday |
| 27028 | 4D96EB78 | Eventuality | Jack Trammell | action | Springfield | 21:22:51 | Monday |
| ... | ... | ... | ... | ... | ... | ... | ... |
| 21799 | 77DBBA7A | Accordion Rhythm | Ahmed Nasr | world | Springfield | 08:50:36 | Friday |
| 6189 | 12A10DD5 | Ekatana | Shantavaani | world | Springfield | 14:30:19 | Friday |
| 18268 | 653E0EF | Kalinka | The Red Army Choirs Of Alexandrov (Les Choeurs... | world | Springfield | 21:08:36 | Wednesday |
| 28665 | D7FB50DA | Drumming Circle | Professor Trance | worldbeat | Springfield | 09:30:47 | Monday |
| 8514 | A439123F | Flip It | unknown | ïîï | Springfield | 09:08:51 | Friday |
42741 rows × 7 columns
# create the shel_general from the df rows,
# where the value in the 'city' column is 'Shelbyville'
shel_general = df[df['city']=='Shelbyville']
shel_general
| user_id | track | artist | genre | city | time | day | |
|---|---|---|---|---|---|---|---|
| 3650 | F10919ED | Ready For The Fire | Valley Of Wolves | acoustic | Shelbyville | 21:33:18 | Monday |
| 14811 | 1C7D9E90 | Prelude – Ne Me Qui Te Pas | unknown | acoustic | Shelbyville | 21:47:42 | Friday |
| 18559 | 897AAB4E | Betcha Never | Cherie | adult | Shelbyville | 09:30:39 | Friday |
| 35110 | 4DCD818F | Blue Skies Again | Jessica Lea Mayfield | adult | Shelbyville | 09:05:48 | Wednesday |
| 49611 | 826C2C51 | Your Song | Ben Cocks | adult | Shelbyville | 20:48:28 | Friday |
| ... | ... | ... | ... | ... | ... | ... | ... |
| 21601 | 6B1890C6 | Unknown Dream | Farhad | world | Shelbyville | 14:46:01 | Wednesday |
| 21666 | F3D7493D | Hello peppinedda my girl josephine | Dany Danubio | world | Shelbyville | 21:16:10 | Monday |
| 41840 | DC347AA4 | La Sitiera | Omara Portuondo | world | Shelbyville | 21:32:36 | Wednesday |
| 50582 | 4DC32D69 | Oath Sign | DJ Mocchi | world | Shelbyville | 14:05:36 | Friday |
| 40493 | AA1730E8 | Anu | Kailash Kokopelli | worldbeat | Shelbyville | 20:16:34 | Wednesday |
18512 rows × 7 columns
# Sorting Genres function
import pandas as pd
pd.to_datetime(arg=df['time'], format='%H:%M:%S')
def genre_weekday(data, day_of_week, first_time, second_time) :
# consecutive filtering
# Create the variable genre_df which will store only those df rows where the day is equal to day=
genre_df = data[data['day'] == day_of_week]
# filter again so that genre_df will store only those rows where the time is smaller than time2=
genre_df = genre_df[genre_df['time'] <= second_time ]
# filter once more so that genre_df will store only rows where the time is greater than time1=
genre_df = genre_df[genre_df['time'] >= first_time]
# group the filtered DataFrame by the column with the names of genres, take the genre column, and find the number of rows for each genre with the count() method
genre_df_count = genre_df.groupby('genre')['user_id'].count()
# sort the result in descending order (so that the most popular genres come first in the Series object)
genre_df_sorted = genre_df_count.sort_values(ascending=False)
# we will return the Series object storing the 15 most popular genres on a given day in a given timeframe
return genre_df_sorted[:15]
Compare the results of the genre_weekday() function for Springfield and Shelbyville on Monday morning (from 7AM to 11AM) and on Friday evening (from 17:00 to 23:00):
# calling the function for Monday morning in Springfield (use spr_general instead of the df table)
print('Monday morning Springfield')
genre_weekday(spr_general, 'Monday', '07:00:00', '11:00:00')
Monday morning Springfield
genre pop 781 dance 549 electronic 480 rock 474 hiphop 286 ruspop 186 world 181 rusrap 175 alternative 164 unknown 161 classical 157 metal 120 jazz 100 folk 97 soundtrack 95 Name: user_id, dtype: int64
# calling the function for Monday morning in Shelbyville (use shel_general instead of the df table)
print('Monday morning Shelbyville')
genre_weekday(shel_general, 'Monday', '07:00', '11:00')
Monday morning Shelbyville
genre pop 218 dance 182 rock 162 electronic 147 hiphop 80 ruspop 64 alternative 58 rusrap 55 jazz 44 classical 40 world 36 rap 32 soundtrack 31 rnb 27 metal 27 Name: user_id, dtype: int64
# calling the function for Friday evening in Springfield
print('Friday evening Springfield')
genre_weekday(spr_general, 'Friday', '17:00', '23:00')
Friday evening Springfield
genre pop 713 rock 517 dance 495 electronic 482 hiphop 273 world 208 ruspop 170 classical 163 alternative 163 rusrap 142 jazz 111 unknown 110 soundtrack 105 rnb 90 metal 88 Name: user_id, dtype: int64
# calling the function for Friday evening in Shelbyville
print('Friday evening Shelbyvile')
genre_weekday(shel_general, 'Friday', '17:00', '23:00')
Friday evening Shelbyvile
genre pop 256 rock 216 electronic 216 dance 210 hiphop 97 alternative 63 jazz 61 classical 60 rusrap 59 world 54 unknown 47 ruspop 47 soundtrack 40 metal 39 rap 36 Name: user_id, dtype: int64
Pop music is the most popular genre in both cities, throughout the week. Dance, Rock, and Electronic are also popular genres throughout the week.
Hypothesis: Shelbyville loves rap music. Springfield's citizens are more into pop.
# count of users listening to various genres
spr_genres = spr_general.groupby('genre')['user_id'].count().sort_values(ascending=False)
# printing the first 10 rows of spr_genres
print('Songs played for each genre : Springfield')
spr_genres.head(10)
Songs played for each genre : Springfield
genre pop 5892 dance 4435 rock 3965 electronic 3786 hiphop 2096 classical 1616 world 1432 alternative 1379 ruspop 1372 rusrap 1161 Name: user_id, dtype: int64
# count of users listening to various genres
shel_genres = shel_general.groupby('genre')['user_id'].count().sort_values(ascending=False)
# printing the first 10 rows from shel_genres
print('Songs played for each genre : Shelbyville')
shel_genres.head(10)
Songs played for each genre : Shelbyville
genre pop 2431 dance 1932 rock 1879 electronic 1736 hiphop 960 alternative 649 classical 646 rusrap 564 ruspop 538 world 515 Name: user_id, dtype: int64
Hypothesis that Springfield citizens are in to pop music is correct. Shelbyville citizens also prefer pop music, while rap music is less popular than hypothesized.