{
"cells": [
{
"cell_type": "markdown",
"metadata": {
"id": "bMcZtJYe9cMJ"
},
"source": [
"# DH-401: Digital Musicology semester project\n",
"---------------\n",
"## Predicting music popularity using DNNs - Milestone 3\n",
"\n",
"### Table of Content\n",
"\n",
"1. [Research Question](#rq)\n",
"1. [Popularity score](#pop)\n",
"1. [Features selection for regression](#feat-selection)\n",
"1. [Regressions](#regs)\n",
"1. [Survey](#survey)\n",
"1. [Explaining the linear regression model of popularity score using audioLIME](#audiolime)\n",
"1. [Appendix: Efforts to use Wave2Vec2.0 for popularity estimation](#wave2vec)\n",
"\n",
"-------------------"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Reasearch Question \n",
"\n",
"It's common in the history of music that some genres and bands gain very high popularity and some aren't noticed at all. It's not really easy to find out which piece will gain an universal acclaim. Or is it? Maybe the popularity of the most known music is the intrinsic feature of that music itself?\n",
"\n",
"In this project, we want to answer the question: to what extent music popularity can be identified solely based on the musical features, and what are the features that make music popular ? Is it possible to identify musical features that play a significant role in songs' popularity ? It's interesting to see whether there are parameters and patterns which lead to the increased music popularity. Needless to say, that this kind of information is priceless for the whole music industry, and for the musicians themselves, as the music which is more popular will, by definition, reach a broader audience and have a higher chance of selling better. \n",
"\n",
"\n",
"To conduct our study, we use the [FMA: A Dataset for Music Analysis](https://arxiv.org/abs/1612.01840) dataset. It is publicly available on [Github](https://github.com/mdeff/fma) and the files are stored on UNIL/SWITCH server.\n",
"\n",
"\n",
"For this study, we are making two main assumptions :\n",
"1. We claim that popular music has a lot of listens. Let's explain that by starting from the definition of \"popular music\". According to Gaynor Jones and Jay Rahn, \"an obvious criterion for a music's popularity is the number of people who experience it: the more people involved, the more popular music\" *(Gaynor Jones and Jay Rahn (1977). Definitions of Popular Music: Recycled. The Journal of Aesthetic Education , Oct., 1977, Vol. 11, No. 4, pp. 79-92, [Jstor](https://www.jstor.org/stable/3332182).)*. That means that a music can be qualified as popular if it reaches a wide audience. However, we shouldn't be too quick with the definition and acknowledge that popularity more complex than that, that \"various groups of people cultivate certain genres within a popular idiom\". *(Gaynor Jones and Jay Rahn (1977). Definitions of Popular Music: Recycled. The Journal of Aesthetic Education , Oct., 1977, Vol. 11, No. 4, pp. 79-92, [Jstor](https://www.jstor.org/stable/3332182).)*. This can be social class, geographical place, race, education, or other. We do not account for any of these in our study. We however try to make our analysis more fine-grained by studying popularity within genres, so that a music is popular if, within its genre, it has a lot of listens.\n",
"\n",
"\n",
"2. There would exist some intrinsic features of each music piece which participate in making it more or less popular than others. We mean by 'intrinsic features', features that are directly calculated from the audio sound wave - and its spectrogram. This approach inscribes our work in the still recent field of Hit Song Science (HSS) whose goal is to \"understand better the relation between intrinsic characteristics of songs and their popularity, regardless of the complex and poorly understood mechanisms of human appreciation and social pressure\" (Pachet, 2012). As a result of this project, we expect to identify several technical characteristics of the music which result in its success.\n",
"\n",
"\n",
"To answer the research question, we are first going to check our two assumptions. For assumption (1), we will conduct a survey in order to get a better intuition on how people are perceiving popular music, and how people like some subset of our dataset. We hope to find that music that people would listen to a lot are the ones that are popular, justifying assumption (1). To check on assumption (2), we implement a Deep Neural Network to see if it manages to predict popularity, because if it does then it's a good indication that there exists musical features that play a role in popularity. After that, we are going to come up with a popularity metric based on the data available. It will take into accound the number of listens of the song, but also of the album and the number of likes and comments. Then we identify discriminating features in music, and apply regression models for musics in each genre to detect the features that play a role in popularity. Finally, the biggest challenge is how we can interpret the features and make sense of them, trying to create listenable interpretations of our model, to answer accuratly our research question.\n",
"\n",
"\n",
"Throughout this study, we will take particular care to pinpoint the limitations of each step that we take. On the survey, we acknowledge that it has some crucial limitations, and we will try to explain the choice of our questions as precisely as possible, and describing how they will help us answer our research question. On the model interpretation, one of the biggest drawback is that the computed features are not really easily interpretable. That's why we try an approach to make it listenable, that we will develop further later in the notebook.\n",
" \n",
" "
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {
"id": "30jJwH7i9cMO"
},
"outputs": [],
"source": [
"from zipfile import ZipFile\n",
"from tqdm.notebook import tqdm\n",
"from pqdm.processes import pqdm\n",
"\n",
"from IPython import display\n",
"\n",
"import librosa\n",
"import soundfile as sf\n",
"import torchaudio\n",
"\n",
"import seaborn as sns\n",
"import pandas as pd\n",
"import numpy as np\n",
"\n",
"import os\n",
"import ast\n",
"\n",
"import statsmodels.api as sm\n",
"import statsmodels.formula.api as smf\n",
"from sklearn.discriminant_analysis import LinearDiscriminantAnalysis\n",
"from sklearn.decomposition import PCA\n",
"\n",
"from scipy.stats import chisquare\n",
"from scipy.stats import shapiro\n",
"\n",
"from nltk import agreement\n",
"\n",
"import matplotlib.pyplot as plt\n",
"from scipy.cluster import hierarchy as hc\n",
"from scipy import stats\n",
"from sklearn import preprocessing\n",
"\n",
"import torch\n",
"import torch.nn as nn\n",
"import torch.nn.functional as F\n",
"\n",
"from transformers import Wav2Vec2PreTrainedModel, Wav2Vec2Model, Wav2Vec2FeatureExtractor\n",
"from transformers.modeling_outputs import SequenceClassifierOutput\n",
"\n",
"from torch.utils.data import DataLoader\n",
"\n",
"from audioLIME.data_provider import DataProvider\n",
"from audioLIME.factorization import DataBasedFactorization\n",
"from audioLIME import lime_audio\n",
"from spleeter.separator import Separator\n",
"\n",
"plt.rcParams['figure.figsize'] = (17, 5)\n",
"\n",
"# You should have fma_large and fma_metadata updated in this directory for the code to run properly\n",
"datasource = \"data\""
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "EZhpKb2x9cMR",
"outputId": "3d278fc5-a67c-4d78-f128-ca14099374d5"
},
"outputs": [
{
"data": {
"text/plain": [
"(106574, 52)"
]
},
"execution_count": 2,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# Load metadata and features.\n",
"# Function based on: https://github.com/mdeff/fma/blob/master/utils.py\n",
"tracks = pd.read_csv(f'{datasource}/fma_metadata/tracks.csv', index_col=0, header=[0, 1])\n",
"\n",
"COLUMNS = [('track', 'tags'), ('album', 'tags'), ('artist', 'tags'),\n",
" ('track', 'genres'), ('track', 'genres_all')]\n",
"for column in COLUMNS:\n",
" tracks[column] = tracks[column].map(ast.literal_eval)\n",
"\n",
"COLUMNS = [('track', 'date_created'), ('track', 'date_recorded'),\n",
" ('album', 'date_created'), ('album', 'date_released'),\n",
" ('artist', 'date_created'), ('artist', 'active_year_begin'),\n",
" ('artist', 'active_year_end')]\n",
"for column in COLUMNS:\n",
" tracks[column] = pd.to_datetime(tracks[column])\n",
"\n",
"SUBSETS = ('small', 'medium', 'large')\n",
"try:\n",
" tracks['set', 'subset'] = tracks['set', 'subset'].astype(\n",
" 'category', categories=SUBSETS, ordered=True)\n",
"except (ValueError, TypeError):\n",
" # the categories and ordered arguments were removed in pandas 0.25\n",
" tracks['set', 'subset'] = tracks['set', 'subset'].astype(\n",
" pd.CategoricalDtype(categories=SUBSETS, ordered=True))\n",
"\n",
"COLUMNS = [('track', 'genre_top'), ('track', 'license'),\n",
" ('album', 'type'), ('album', 'information'),\n",
" ('artist', 'bio')]\n",
"for column in COLUMNS:\n",
" tracks[column] = tracks[column].astype('category')\n",
"\n",
"\n",
"tracks.shape"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "ZqbOkcuM9cMW"
},
"source": [
"## Popularity score \n",
"\n",
"To define popularity, we will use a score, instead of 'popular' or 'unpopular', as it is more accurate to talk about levels of popularity, rather than a binary classification. *(Russel B. Nye, \"Notes for an Introduction to a Discussion of Popular Culture, Journal of Popular Culture, Vol. 4 (Spring 1971):1031-38, on the need to recognize degrees of popularity)*.\n",
"\n",
"\n",
"The dataset contains different measures that can be used to define popularity: number of listens, likes and comments. These features are present at two levels (song and album). The two levels come from Free Music Archive's API and both encompass important information concerning the dispersion of the songs. Thus, we first build a DataFrame containing all these scores for all of the songs available in the dataset."
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "HhpHdEJN9cMY"
},
"source": [
"### Popularity features "
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {
"id": "vHuKsOS59cMZ"
},
"outputs": [],
"source": [
"POP_FEATURES = [\"listens\", \"favorites\", \"comments\"]"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "LHX6kQ3p9cMb",
"outputId": "62ef888a-2341-4a70-d8d0-7049f72701cb"
},
"outputs": [
{
"data": {
"text/html": [
"
\n",
"\n",
"
\n",
" \n",
"
\n",
"
\n",
"
date_created
\n",
"
genre_top
\n",
"
listens
\n",
"
favorites
\n",
"
comments
\n",
"
album_listens
\n",
"
album_favorites
\n",
"
album_comments
\n",
"
\n",
"
\n",
"
track_id
\n",
"
\n",
"
\n",
"
\n",
"
\n",
"
\n",
"
\n",
"
\n",
"
\n",
"
\n",
" \n",
" \n",
"
\n",
"
124070
\n",
"
2015-08-24 16:32:25
\n",
"
NaN
\n",
"
353
\n",
"
1
\n",
"
0
\n",
"
11139
\n",
"
0
\n",
"
0
\n",
"
\n",
"
\n",
"
48838
\n",
"
2011-06-11 15:47:29
\n",
"
NaN
\n",
"
444
\n",
"
0
\n",
"
0
\n",
"
6043
\n",
"
1
\n",
"
3
\n",
"
\n",
"
\n",
"
89888
\n",
"
2013-08-30 16:15:29
\n",
"
NaN
\n",
"
32669
\n",
"
18
\n",
"
0
\n",
"
65976
\n",
"
1
\n",
"
0
\n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" date_created genre_top listens favorites comments \\\n",
"track_id \n",
"124070 2015-08-24 16:32:25 NaN 353 1 0 \n",
"48838 2011-06-11 15:47:29 NaN 444 0 0 \n",
"89888 2013-08-30 16:15:29 NaN 32669 18 0 \n",
"\n",
" album_listens album_favorites album_comments \n",
"track_id \n",
"124070 11139 0 0 \n",
"48838 6043 1 3 \n",
"89888 65976 1 0 "
]
},
"execution_count": 4,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"pop_df = tracks[\"track\"][[\"date_created\", \"genre_top\"]+POP_FEATURES]\n",
"\n",
"# add informations collected at the \"album level\"\n",
"for pop_feat in POP_FEATURES:\n",
" pop_df[\"album_\"+pop_feat] = [count for count in tracks[\"album\"][pop_feat]]\n",
"\n",
"pop_df.sample(3)"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "O7yRsc-Z9cMd"
},
"source": [
"The number of listens per song and per album reflect different realities, in that they are not directly based on one another (the number of listens of an album is not the sum of the number of listens of its songs and the number of listens of the songs don't take into account the number of listens of the album). In other words, if $l_s$ is the number of listens of a song, $a$ refers to albums and $\\{s\\in a\\}$ the set of songs in one album, $$ \\sum_{s\\in a}l_s \\neq l_a $$\n",
"\n",
"This fact is illustrated below and two extreme examples are underlined."
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 66,
"referenced_widgets": [
"44ec8feff73d4e6abd372e5e88c79c76",
"2909cd6e3ee844bbaf818d7f98007288",
"fd2983b086e94940a10b5d7d51fd4530",
"bb42199b84314b7995f56e39c850c25c",
"26e1ad0b929d4916876da4db38bf1869",
"491b8e5ab24548e38c65cbbfed26e091",
"cda6871a5a5c4625939f7a72040cbd35",
"a6d4c1db736047ef98ffc7f9d93be588"
]
},
"id": "ATAv494i9cMe",
"outputId": "4b54bcf7-0956-403f-f40c-af8811deae43"
},
"outputs": [
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "1dea96c0fc1f47319a99a167db5439b3",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
"HBox(children=(FloatProgress(value=0.0, max=14395.0), HTML(value='')))"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"\n"
]
}
],
"source": [
"# build a DataFrame of albums\n",
"albums = tracks[\"album\"].groupby([\"id\"]).mean()[[\"listens\"]]\n",
"# remove songs without album\n",
"albums = albums[albums.listens > 0]\n",
"\n",
"albums[\"sum_songs\"] = 0\n",
"albums[\"album\"] = 0\n",
"albums[\"songs_minus_album\"] = 0\n",
"albums[\"songs_ratio_album\"] = 0\n",
"\n",
"for alb_id, _ in tqdm(albums.iterrows(), total=albums.shape[0]):\n",
" # retrieve track ids for the album\n",
" curr_track_ids = tracks[\"album\"][tracks[\"album\"].id == alb_id].index.values\n",
" # compute sum of the listens of the album's songs\n",
" curr_sum_songs_listens = pop_df.loc[curr_track_ids].listens.sum()\n",
" # retrieve album's number of listens\n",
" curr_album_listens = pop_df.loc[curr_track_ids].album_listens.iloc[0]\n",
" \n",
" # add these information to the `albums` DataFrame\n",
" albums.loc[alb_id, \"sum_songs\"] = curr_sum_songs_listens\n",
" albums.loc[alb_id, \"album\"] = curr_album_listens\n",
" # compute difference and ratio\n",
" albums.loc[alb_id, \"songs_minus_album\"] = curr_sum_songs_listens-curr_album_listens\n",
" albums.loc[alb_id, \"songs_ratio_album\"] = curr_sum_songs_listens/curr_album_listens\n",
" \n",
"albums = albums.sort_values(by=\"songs_minus_album\")"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 357
},
"id": "6QCR2DOV9cMi",
"outputId": "b0ffde61-bc36-4dfc-a4d2-c7b3f7d4fcd5"
},
"outputs": [
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAA9wAAAFUCAYAAADWLDSBAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8li6FKAAAcV0lEQVR4nO3df5Dnd10f8OerCQHM0QUET5ukvTBJIykRCztBmWovgngYT5CJmpQiscFTOoztTGo9x44dp9MWq3RGazrMiRmqtRwUrSYkNiJ4Um1GQiQ0iTF6xjAc0ESMLBxSwsGrf+wX3Wzusru33/d+d7/3eMx85vbz/n5+vHbnfZ+9533en/enujsAAADAdP2NWRcAAAAA80jgBgAAgAEEbgAAABhA4AYAAIABBG4AAAAYQOAGAACAAc6edQFJ8qxnPav37Nkz6zJgUz7zmc/k3HPPnXUZMBX6M/NEf2ae6M/Mk3nqz3feeecnuvvZq9u3ReDes2dPPvCBD8y6DNiUI0eOZO/evbMuA6ZCf2ae6M/ME/2ZeTJP/bmqPnyydkPKAQAAYICZBu6q2l9Vh5aWlmZZBgAAAEzdTAN3d9/c3QcWFhZmWQYAAABMnSHlAAAAMIDADQAAAAMI3AAAADCAwA0AAAADCNwAAAAwgMANAAAAAwjcAAAAMIDADQAAAAOcPesCAJi9PQdvecz69ZedyLUr2h5845VbXRIAwI7nDjcAAAAMIHADAADAAAI3AAAADDAkcFfVuVV1Z1V924jjAwAAwHa3rsBdVTdW1cNVdc+q9n1VdX9VHa2qgys++uEk75hmoQAAALCTrPcO91uT7FvZUFVnJbkhycuTXJrkmqq6tKpemuQPkjw0xToBAABgR1nXa8G6+31VtWdV8+VJjnb3A0lSVYeTvCLJriTnZjmEf7aqbu3uL06tYgAAANgBqrvXt+Fy4H5Xdz9vsn5Vkn3d/brJ+muSvKi73zBZvzbJJ7r7Xac43oEkB5Jk9+7dLzx8+PCmvhGYtePHj2fXrl2zLgNOy90fXXrM+u6nJg999q/XLztvYYsrgulxfWae6M/Mk3nqz1dcccWd3b24un1dd7hPoU7S9lfpvbvf+kQ7d/ehJIeSZHFxsffu3buJUmD2jhw5Ev2Ynerag7c8Zv36y07kTXf/9a+IB1+9d4srgulxfWae6M/MkzOhP29mlvJjSS5YsX5+ko9trhwAAACYD5sJ3HckubiqLqyqc5JcneSmjRygqvZX1aGlpaW1NwYAAIAdZL2vBXtbktuTXFJVx6rquu4+keQNSW5Lcl+Sd3T3vRs5eXff3N0HFhY8GwgAAMB8We8s5decov3WJLdOtSIAAACYA5sZUr5phpQDAAAwr2YauA0pBwAAYF7NNHADAADAvDKkHAAAAAYwpBwAAAAGMKQcAAAABhC4AQAAYADPcAMAAMAAnuEGAACAAQwpBwAAgAEEbgAAABhA4AYAAIABTJoGAAAAA5g0DQAAAAYwpBwAAAAGELgBAABgAIEbAAAABhC4AQAAYACzlAMAAMAAZikHAACAAQwpBwAAgAEEbgAAABhA4AYAAIABBG4AAAAYQOAGAACAAbwWDAAAAAbwWjAAAAAYwJByAAAAGEDgBgAAgAEEbgAAABhA4AYAAIABBG4AAAAYQOAGAACAAQRuAAAAGEDgBgAAgAEEbgAAABhgpoG7qvZX1aGlpaVZlgEAAABTN9PA3d03d/eBhYWFWZYBAAAAU2dIOQAAAAwgcAMAAMAAAjcAAAAMIHADAADAAAI3AAAADCBwAwAAwAACNwAAAAwgcAMAAMAAAjcAAAAMIHADAADAAAI3AAAADDD1wF1Vz62qN1fVO6vq9dM+PgAAAOwE6wrcVXVjVT1cVfesat9XVfdX1dGqOpgk3X1fd/9Aku9Ksjj9kgEAAGD7W+8d7rcm2beyoarOSnJDkpcnuTTJNVV16eSzb0/yO0neM7VKAQAAYAdZV+Du7vcleWRV8+VJjnb3A939aJLDSV4x2f6m7n5xkldPs1gAAADYKc7exL7nJfnIivVjSV5UVXuTvCrJk5PcuonjAwAAwI61mcBdJ2nr7j6S5MiaO1cdSHIgSXbv3p0jR9bcBba148eP68fsWNdfduIx67uf+tg2fZudzPWZeaI/M0/OhP68mcB9LMkFK9bPT/Kx9e7c3YeSHEqSxcXF3rt37yZKgdk7cuRI9GN2qmsP3vKY9esvO5E33f3XvyIefPXeLa4Ipsf1mXmiPzNPzoT+vJnXgt2R5OKqurCqzklydZKbNnKAqtpfVYeWlpY2UQYAAABsP+t9Ldjbktye5JKqOlZV13X3iSRvSHJbkvuSvKO7793Iybv75u4+sLCwsNG6AQAAYFtb15Dy7r7mFO23xsRoAAAA8DibGVK+aYaUAwAAMK9mGrgNKQcAAGBezTRwAwAAwLwSuAEAAGAAz3ADAADAAJ7hBgAAgAEMKQcAAIABBG4AAAAYwDPcAAAAMIBnuAEAAGAAQ8oBAABgAIEbAAAABvAMNwAAAAzgGW4AAAAYwJByAAAAGEDgBgAAgAEEbgAAABhA4AYAAIABzFIOAAAAA5ilHAAAAAYwpBwAAAAGELgBAABgAIEbAAAABhC4AQAAYACBGwAAAAY4e9YFALD97Tl4y5rbPPjGK7egEgCAncN7uAEAAGAA7+EGAACAATzDDQAAAAMI3AAAADCAwA0AAAADCNwAAAAwgMANAAAAAwjcAAAAMIDADQAAAAMI3AAAADDATAN3Ve2vqkNLS0uzLAMAAACmbqaBu7tv7u4DCwsLsywDAAAAps6QcgAAABhA4AYAAIABBG4AAAAYQOAGAACAAQRuAAAAGEDgBgAAgAEEbgAAABhA4AYAAIABBG4AAAAYQOAGAACAAQRuAAAAGEDgBgAAgAGGBO6qemVV/VxV/VpVvWzEOQAAAGA7W3fgrqobq+rhqrpnVfu+qrq/qo5W1cEk6e5f7e7vS3Jtku+easUAAACwA2zkDvdbk+xb2VBVZyW5IcnLk1ya5JqqunTFJv9q8jkAAACcUdYduLv7fUkeWdV8eZKj3f1Adz+a5HCSV9Syn0jy6939+9MrFwAAAHaG6u71b1y1J8m7uvt5k/Wrkuzr7tdN1l+T5EVJ/ijJa5PckeSu7n7zSY51IMmBJNm9e/cLDx8+vKlvBGbt+PHj2bVr16zLgNNy90eXHrO++6nJQ5/d2DEuO29hihXB9Lg+M0/0Z+bJPPXnK6644s7uXlzdfvYmj1snaevu/pkkP/NEO3b3oSSHkmRxcbH37t27yVJgto4cORL9mJ3q2oO3PGb9+stO5E13b+xXxIOv3jvFimB6XJ+ZJ/oz8+RM6M+bnaX8WJILVqyfn+RjmzwmAAAA7HibDdx3JLm4qi6sqnOSXJ3kpvXuXFX7q+rQ0tLS2hsDAADADrKR14K9LcntSS6pqmNVdV13n0jyhiS3JbkvyTu6+971HrO7b+7uAwsLnvsDAABgvqz7Ab3uvuYU7bcmuXVqFQEAAMAc2OykaZtSVfuT7L/oootmWQbAXNuzakI0AAC2xmaf4d4UQ8oBAACYVzMN3AAAADCvZhq4zVIOAADAvJrpM9zdfXOSmxcXF79vlnUAsHlrPSv+4Buv3KJKAAC2B0PKAQAAYACBGwAAAAbwDDcAAAAM4LVgAAAAMIAh5QAAADCAwA0AAAADCNwAAAAwgEnTAAAAYACTpgEAAMAAhpQDAADAAAI3AAAADCBwAwAAwAACNwAAAAxglnIAAAAYwCzlAAAAMIAh5QAAADCAwA0AAAADCNwAAAAwgMANAAAAA5w96wIAODPsOXjLmts8+MYrt6ASAICt4bVgAAAAMIDXggEAAMAAhpQD7GDrGaYNAMBsmDQNAAAABhC4AQAAYACBGwAAAAYQuAEAAGAAk6YBsG14VzcAME/c4QYAAIABBG4AAAAYQOAGAACAAWYauKtqf1UdWlpammUZAAAAMHUzDdzdfXN3H1hYWJhlGQAAADB1hpQDAADAAAI3AAAADCBwAwAAwAACNwAAAAwgcAMAAMAAAjcAAAAMIHADAADAAAI3AAAADCBwAwAAwAACNwAAAAwgcAMAAMAAAjcAAAAMMPXAXVXPqaqfr6p3TvvYAAAAsFOsK3BX1Y1V9XBV3bOqfV9V3V9VR6vqYJJ09wPdfd2IYgEAAGCnWO8d7rcm2beyoarOSnJDkpcnuTTJNVV16VSrAwAAgB1qXYG7u9+X5JFVzZcnOTq5o/1oksNJXjHl+gAAAGBHqu5e34ZVe5K8q7ufN1m/Ksm+7n7dZP01SV6U5F8n+bdJvjnJW7r735/ieAeSHEiS3bt3v/Dw4cOb+kZg1o4fP55du3bNugx2kLs/ujTrEk5p91OThz476ypO7rLzFmZdAjuM6zPzRH9mnsxTf77iiivu7O7F1e1nb+KYdZK27u4/T/IDa+3c3YeSHEqSxcXF3rt37yZKgdk7cuRI9GM24tqDt8y6hFO6/rITedPdm/kVMc6Dr9476xLYYVyfmSf6M/PkTOjPm5ml/FiSC1asn5/kY5srBwAAAObDZgL3HUkurqoLq+qcJFcnuWkjB6iq/VV1aGlp+w6rBAAAgNOx3teCvS3J7UkuqapjVXVdd59I8oYktyW5L8k7uvvejZy8u2/u7gMLC57HAwAAYL6s6wG97r7mFO23Jrl1qhUBAADAHJjpjDhVtT/J/osuumiWZQBs2J41Jjx78I1XblElAABsV5t5hnvTDCkHAABgXs00cAMAAMC8mmngNks5AAAA88qQcgAAABjAkHIAAAAYQOAGAACAAbwWDIAdZRqvZFvrGOs5zjSOAQDMN89wAwAAwACGlAMAAMAAAjcAAAAMIHADAADAACZNA4CTWM+kaJs9hknVAGC+mTQNAAAABjCkHAAAAAYQuAEAAGAAgRsAAAAGELgBAABgALOUAzvGemaNNusz05hdHABgGsxSDgAAAAMYUg4AAAADCNwAAAAwgMANAAAAAwjcAAAAMIDADQAAAAN4LRjAAF5NxXbilXoAMBteCwYAAAADGFIOAAAAAwjcAAAAMIDADQAAAAMI3AAAADCAwA0AAAADCNwAAAAwgMANAAAAAwjcAAAAMIDADQAAAAOcPcuTV9X+JPsvuuiiWZYBrGHPwVvW3ObBN165BZUAo6zn7/la1nMdWOs8riUAzJOZ3uHu7pu7+8DCwsIsywAAAICpM6QcAAAABhC4AQAAYACBGwAAAAYQuAEAAGAAgRsAAAAGELgBAABgAIEbAAAABhC4AQAAYACBGwAAAAYQuAEAAGAAgRsAAAAGELgBAABggLOnfcCqOjfJf07yaJIj3f1L0z4HAAAAbHfrusNdVTdW1cNVdc+q9n1VdX9VHa2qg5PmVyV5Z3d/X5Jvn3K9AAAAsCOsd0j5W5PsW9lQVWcluSHJy5NcmuSaqro0yflJPjLZ7AvTKRMAAAB2lnUF7u5+X5JHVjVfnuRodz/Q3Y8mOZzkFUmOZTl0r/v4AAAAMG+qu9e3YdWeJO/q7udN1q9Ksq+7XzdZf02SFyX54SQ/m+T/JfmdUz3DXVUHkhxIkt27d7/w8OHDm/pGtsrdH11ac5vLzlvYgkrYbo4fP55du3ate/u1+tK0+tE0zrOefj8Na9Uyrb9/O+lnMiu7n5o89NlZVzH/turv+VbZqr87Gz3P6fbneft9vlW/dzZbRzJ/P/tp2ui/NxhPn3689V7r17o+76Sf2xVXXHFndy+ubt/MpGl1krbu7s8k+d61du7uQ0kOJcni4mLv3bt3E6VsnWsP3rLmNg++eu/4Qth2jhw5ko3047X60rT60TTOs55+Pw1r1TKtv3876WcyK9dfdiJvunvq82qyylb9Pd8qW/V3Z6PnOd3+PG+/z7fq985m60jm72c/TRv99wbj6dOPt95r/VrX53n4uW1myPexJBesWD8/ycc2coCq2l9Vh5aWtsf/zAMAAMC0bCZw35Hk4qq6sKrOSXJ1kps2coDuvrm7Dyws7JyhAgAAALAe630t2NuS3J7kkqo6VlXXdfeJJG9IcluS+5K8o7vvHVcqAAAA7BzreqCpu685RfutSW493ZNX1f4k+y+66KLTPQQAAABsSzN9bZch5QAAAMwr78kGAACAAQRuAAAAGGCmgdtrwQAAAJhXnuEGAACAAQwpBwAAgAGqu2ddQ6rqz5J8eNZ1wCY9K8knZl0ETIn+zDzRn5kn+jPzZJ7689/p7mevbtwWgRvmQVV9oLsXZ10HTIP+zDzRn5kn+jPz5Ezoz4aUAwAAwAACNwAAAAwgcMP0HJp1ATBF+jPzRH9mnujPzJO578+e4QYAAIAB3OEGAACAAQRuWKGqvrOq7q2qL1bV4qrPfqSqjlbV/VX1LSvaX1hVd08++5mqqkn7k6vq7ZP236uqPSv2eW1V/fFkee2K9gsn2/7xZN9zxn/X8FhVtW/Sz49W1cFZ18OZq6purKqHq+qeFW3PrKp3T66T766qZ6z4bPh1Gk5XVV1QVb9VVfdN/q3xzybt+jQ7TlU9pareX1UfmvTnH5+068+rdbfFYpksSZ6b5JIkR5Isrmi/NMmHkjw5yYVJ/iTJWZPP3p/k65NUkl9P8vJJ+z9N8ubJ11cnefvk62cmeWDy5zMmXz9j8tk7klw9+frNSV4/65+J5cxakpw16d/PSXLOpN9fOuu6LGfmkuQbk7wgyT0r2v5DkoOTrw8m+YnJ11tynbZYTndJ8lVJXjD5+mlJ/mjSb/Vpy45bJn1v1+TrJyX5vSRfpz8/fnGHG1bo7vu6+/6TfPSKJIe7+3Pd/adJjia5vKq+Ksnf7O7be/kK8AtJXrlin/8y+fqdSV4y+R+7b0ny7u5+pLv/Ism7k+ybfPZNk20z2fdLx4KtcnmSo939QHc/muRwlvsybLnufl+SR1Y1r7y2rrxODr9OT/875EzS3R/v7t+ffP3pJPclOS/6NDtQLzs+WX3SZOnoz48jcMP6nJfkIyvWj03azpt8vbr9Mft094kkS0m+/AmO9eVJPjnZdvWxYKucqn/CdrG7uz+eLAeYJF8xad+K6zRMxWRo7N/P8l1BfZodqarOqqq7kjyc5QCsP5/E2bMuALZaVf1mkq88yUc/2t2/dqrdTtLWT9B+Ovs80bFgq+iH7FRbcZ2GTauqXUl+Ock/7+5PTR5XPemmJ2nTp9k2uvsLSb62qp6e5H9U1fOeYPMztj+7w80Zp7tf2t3PO8lyqrCdLP/P2QUr1s9P8rFJ+/knaX/MPlV1dpKFLA+NPNWxPpHk6ZNtVx8Ltsqp+idsFw9NhiBm8ufDk/atuE7DplTVk7Ictn+pu39l0qxPs6N19yezPP/RvujPjyNww/rclOTqyWyJFya5OMn7J0NlPl1VXzd5puR7kvzain2+NGviVUneO3k25bYkL6uqZ0xmbnxZktsmn/3WZNtM9n2i/wSAEe5IcnEtz5h/TpYnKblpxjXBSiuvrSuvk8Ov06O/MebbpP/9fJL7uvs/rvhIn2bHqapnT+5sp6qemuSlSf4w+vPjzXrWNotlOy1JviPL/2v2uSQPZTkIf+mzH83yjIr3ZzJ74qR9Mck9k89+NklN2p+S5L9neVKI9yd5zop9/smk/WiS713R/pzJtkcn+z551j8Ty5m3JPnWLM+e+ydZftRi5jVZzswlyduSfDzJ5yfX5uuy/Pzee5L88eTPZ67Yfvh12mI53SXJP8jysNf/k+SuyfKt+rRlJy5JvibJByf9+Z4kPzZp159XLV/6ZgAAAIApMqQcAAAABhC4AQAAYACBGwAAAAYQuAEAAGAAgRsAAAAGELgBAABgAIEbAAAABhC4AWBGquoNVdUbXN4y67oBgPURuAFgdk7MugAAYJzq7lnXAABnpKp6SpJbk1yR5FNJXtLdH5htVQDAtAjcADBDVXVukt9I8uIkjyTZ2913z7YqAGAaBG4AmLGqWkjyniQvTPJQkn/Y3ffPtioAYLM8ww0AA9SyD1XVa9fatruXknxLkruT7E7ym1W15zTPe3VVPVpVT1pjuxuq6udP5xzTrgUA5pXADQBjfFeSZyT5b+vZuLv/PMk3J7k/yflJ3ltV553GeZ+f5A+6+/NrbPeTSV5dVRedxjmmXQsAzCWBGwDG+MEkv7iRsNndDyV5SZI/TXJhkvdU1Vds8Lxfm+RD6zjXg0l+J8nrN3j8qdcCAPNK4AaAKZvcNX5xkndudN/u/miSb0pyLMklSd5dVc/cwCGen+SudW77y1m+yz3q3wMbqQUA5o7ADQDT95Ikn8lp3t2d3H1+82T1a7I81HxNVfXsJF+1gfP+7yw/M37ZBkscUQsAzB2BGwCm74VJ7uvuL57OzlX1yiQ/Pln90e5++zp3/drJn+sNufcm+UKSyzdQ3npttBYAmDtnz7oAAJhDX5nkE6ezY1V9U5LDSc5K8qbu/ncb2P35SY5NJmBbU3efqKpPZrne1XUsZPkO9VrH+MNp1AIA80jgBoDpe0qSv9zoTlW1mORXkzw5yY3d/S82eIjnZ+N3lD+X5XpX+84kP7eO/WuKtQDAXDGkHACm75EkT9/IDlX11Ul+PcnTkvxKkgOncd7TmRX86Vmu9zG6+y3dXWstU64FAOaKwA0A03d/ll/rtS5V9beT/EaSZyX5zST/qLu/sJETVtU5Sb46q2YFr6pXVdUdVXVXVd1XVc9Z8dmzk3xZkj/ayLlOp5aqellV3V5VH6yqe6vqG6Z5TgDYjgwpB4Dp+90kP1ZVz+7uP3uiDavqWVkO2xck+b0k39HdnzuNc/69LP9e/6u7ypPg+9NJntvdx6tqV5aHkH/JYpLO8mzl0/S4WpLcmeTF3d1V9dIkP5Tkf035vACwrbjDDQDTdyTLw7T3PdFGVfW0JP8zy+/bvjfJt3b38dM85/Oz/Nz40RVtneSzSW6YhNy/7O7Pr/h8X5LfHjCx2clquSrJ71bVh5IcmtQFAHOtunvWNQDA3Kmqn05yUXdf+QTb/GKSf7zBQ3+4u/dsoI6nJHlZktcn+bPu/p5J+1lJPpzkYHf/1w3WsCFVdVWS67I8VP4vquqnkjyywRnYAWDHcYcbAMb4ySR7q+rvPsE2a752azOq6pIkn+vum5LcmOXntb/kO7N8l/nwyBomnp/k/ZOw/YIk35/kg1twXgCYKXe4AWCQqro6yce7+7dndP63JPnGJJ9O8lCS7+/uj0w+uybJR7v7fVtQx8VJbk7yqSTvTfLdSb6+u//v6HMDwCwJ3AAAADCAWcoBYJuoqguT/Msszx7+1BUffW933zGbqgCA0+UONwBsA1X1vCQ3JfmRJDd1t1m8AWCHM2kaAGyBqnpVVd1RVXdV1X1V9ZxVm/ynJK/v7revFbar6mVVdXtVfbCq7q2qbxhXOQBwutzhBoDBquqcJH+S5LndfbyqdmV59vDPTz7/siR3dfcTzWi+8nhfnuXXavXk/do/2N3fPqp+AOD0eIYbAMbrLL+C64bJu7ff291fXPH52UnOr6p7TrH/6me4r0ry2qo6N8nTkni+GwC2IYEbAAbr7s9X1dckeVmS65N8z2T50uefqqrfT/Jt3f3JJzpWVV2V5JVJrpy81/qnkjwyrnoA4HR5hhsABquqS7I8hPymJDcm+bKTbPZvkry9qv7WGod7fpL3T8L2C5J8f5IPTrVgAGAq3OEGgPF+KMk3VtWnkzyU5ZD8GN19W1V9IcnhqnpGklrx8coh5b+Q5OaqenmS9yb5RARuANiWTJoGAAAAAxhSDgAAAAMI3AAAADCAwA0AAAADCNwAAAAwgMANAAAAAwjcAAAAMIDADQAAAAMI3AAAADDA/wdG/XBbQOWrWgAAAABJRU5ErkJggg==\n",
"text/plain": [
"
"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"# histogram of the sum of album's songs' listens - album's listens\n",
"albums[\"songs_minus_album\"].hist(bins=100)\n",
"plt.xlabel(r\"$(\\sum_{s\\in a}l_s) - l_a$\", fontsize=15)\n",
"plt.semilogy()\n",
"plt.show()"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "gQsgVU6b9cMk",
"outputId": "2750aad1-b0be-4e61-9a03-48cc1eb0f58a"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"The album 'Music For Media Vol. 3' containing 8 songs was listened 247274 times as an album, but individually the songs account for less listens: 72202.\n"
]
}
],
"source": [
"id_album = albums[\"songs_minus_album\"].idxmin() #10953\n",
"\n",
"id_tracks = tracks[\"album\"][tracks[\"album\"].id == id_album].index.values\n",
"\n",
"print(\"The album '{album_title}' containing {n_songs} song{plural} was listened {alb_listens} times as an album, \\\n",
"but individually the song{plural} account{sing} for {more_or_less} listens: {songs_sum_listens}.\"\\\n",
" .format(album_title=tracks[\"album\"].loc[id_tracks].iloc[0].title\n",
" , n_songs=len(id_tracks)\n",
" , plural='s' if len(id_tracks)>1 else ''\n",
" , sing='' if len(id_tracks)>1 else 's'\n",
" , alb_listens=pop_df.loc[id_tracks].iloc[0].album_listens\n",
" , songs_sum_listens=np.sum(pop_df.loc[id_tracks].listens)\n",
" , more_or_less='more' \n",
" if (np.sum(pop_df.loc[id_tracks].listens)/pop_df.loc[id_tracks].iloc[0].album_listens)>1 \n",
" else 'less'\n",
" )\n",
" )"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "8YO8XyAb9cMp",
"outputId": "8383969c-f5f7-4769-bac4-05361153ef01"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"The album '...Plays Guitar' containing 7 songs was listened 386403 times as an album, but individually the songs account for more listens: 690479.\n"
]
}
],
"source": [
"id_album = albums[\"songs_minus_album\"].idxmax() #7690\n",
"\n",
"id_tracks = tracks[\"album\"][tracks[\"album\"].id == id_album].index.values\n",
"\n",
"print(\"The album '{album_title}' containing {n_songs} song{plural} was listened {alb_listens} times as an album, \\\n",
"but individually the song{plural} account{sing} for {more_or_less} listens: {songs_sum_listens}.\"\\\n",
" .format(album_title=tracks[\"album\"].loc[id_tracks].iloc[0].title\n",
" , n_songs=len(id_tracks)\n",
" , plural='s' if len(id_tracks)>1 else ''\n",
" , sing='' if len(id_tracks)>1 else 's'\n",
" , alb_listens=pop_df.loc[id_tracks].iloc[0].album_listens\n",
" , songs_sum_listens=np.sum(pop_df.loc[id_tracks].listens)\n",
" , more_or_less='more' \n",
" if (np.sum(pop_df.loc[id_tracks].listens)/pop_df.loc[id_tracks].iloc[0].album_listens)>1 \n",
" else 'less'\n",
" )\n",
" )"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "wm8wPEQD9cMq"
},
"source": [
"### Removing songs without album \n",
"\n",
"The number of listens, likes and comments of albums bring supplementary, non negligible, information about the public a song could have reached. However some songs don't appear in albums (reflected by $-1$ for album listens, comments, and favorites counts in the DataFrame). These songs represent appproximately 3.3% of the dataset, so they are left aside to be able to compute a more coherent popularity score based on both of the disposable levels."
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "RCH-_Krr9cMr",
"outputId": "fce92e01-7e80-41bc-ed8f-6c48d55a5cea"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"3.31% songs have no album.\n"
]
}
],
"source": [
"# songs without albums have -1 as count for listens, favorites, comments\n",
"no_album = pop_df.album_listens < 0\n",
"print(\"{0:.2f}% songs have no album.\".format(100*np.sum(no_album)/len(pop_df)))"
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/"
},
"id": "_J3PcyWt9cMs",
"outputId": "5920e4b8-71b4-4895-e249-7371450d016e"
},
"outputs": [
{
"data": {
"text/plain": [
"(103045, 8)"
]
},
"execution_count": 10,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# keeping only songs with album informations\n",
"pop_df = pop_df[~no_album].copy()\n",
"pop_df.shape"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "qni9og_v9cMs"
},
"source": [
"The distribution of the logarithmic transform of the different features are plotted below, showing that they reflect different realities and justifying that the album level should not be neglected, as it appears to be largely used by FMA's users."
]
},
{
"cell_type": "code",
"execution_count": 11,
"metadata": {
"id": "1tJ5JM9o9cMt"
},
"outputs": [],
"source": [
"def log_transform(serie):\n",
" \"\"\"log transform for serie of non-negative values\"\"\"\n",
" return np.log((serie)+1)"
]
},
{
"cell_type": "code",
"execution_count": 12,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 338
},
"id": "73AjiWj69cMt",
"outputId": "cefbe976-0577-4929-f852-ecf627cf0209"
},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"
"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"f, axes = plt.subplots(nrows=1,ncols=3, sharex=False)\n",
"\n",
"for ax, feat in zip(axes, POP_FEATURES) :\n",
" ax.hist(log_transform(pop_df[\"album_\"+feat]), bins=100, alpha=0.6, label=\"album\")\n",
" ax.hist(log_transform(pop_df[feat]), bins=100, alpha=0.6, label=\"song\")\n",
" ax.set_xlabel(r\"$\\log(1+$\"+feat+\"$)$\")\n",
" ax.legend()\n",
" ax.semilogy()\n",
"plt.legend()\n",
"plt.show()"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "rKs5-i889cMv"
},
"source": [
"### Computing popularity scores \n",
"\n",
"The popularity score can then be computed as the one-dimensional projection of the logarithmic transformation of these features using PCA to encompass the maximum informations of these different dimensions in a single value per song.\n",
"\n",
"This popularity score can then be thought of as a proxy of the ability of a song to reach a great audience. The number of likes and comments add informations about the engagement of publics towards the songs.\n",
"It could potentially embrace social dynamics, an hypothesis could be that users that have liked or commented a song are more prone to recommend or play this song to their entourage, thus spreading the song."
]
},
{
"cell_type": "code",
"execution_count": 13,
"metadata": {
"id": "GC5wOxsK9cMv"
},
"outputs": [],
"source": [
"# without `interest` but with album scores\n",
"X = pop_df[POP_FEATURES+[\"album_\"+pop_feat for pop_feat in POP_FEATURES]].copy()\n",
"\n",
"for pop_feat in X.columns:\n",
" # using log transform for the pca\n",
" X[pop_feat] = log_transform(X[pop_feat])\n",
"\n",
"pca = PCA(n_components=1, svd_solver=\"full\").fit(X)\n",
"#computing popularity score as a 1D projection of the popularity features\n",
"pop_df[\"pop_score\"] = pca.transform(X) "
]
},
{
"cell_type": "code",
"execution_count": 14,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 321
},
"id": "rIQyHDGZ9cMv",
"outputId": "2e0ecff0-ce06-43be-d467-63615ad780c6"
},
"outputs": [
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAA9oAAAEwCAYAAABIVwmKAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy8li6FKAAAcqUlEQVR4nO3dfbhlV10f8O+PCVEQhEJGhbxNxAANr8KQSoEaFHgCtAaUmgCKiBqjRkSLklpqw2NbglipYmSMmIKCpsqbAwRCpQQsEJlEQt4gdBqiGYIQ0IIoJQn8+sfe45xcbmZOkjW592Y+n+c5z90v6+y97pw1+5zvWuvsW90dAAAAYIw7rXUFAAAA4I5E0AYAAICBBG0AAAAYSNAGAACAgQRtAAAAGEjQBgAAgIEOWqsTH3LIIb1ly5a1Oj0AAADcahdddNFnu3vzavvWLGhv2bIlF1544VqdHgAAAG61qvrLm9tn6jgAAAAMJGgDAADAQII2AAAADCRoAwAAwECCNgAAAAwkaAMAAMBAgjYAAAAMJGgDAADAQII2AAAADCRoAwAAwECCNgAAAAx00FpXYCPYctrb17oK7CdXn/HUta4CAABwB2NEGwAAAAYStAEAAGAgQRsAAAAG8h1tAOAm3Jvkjsu9SQBuH0a0AQAAYCBBGwAAAAYStAEAAGAgQRsAAAAGErQBAABgIEEbAAAABhK0AQAAYCBBGwAAAAYStAEAAGAgQRsAAAAGErQBAABgIEEbAAAABloqaFfV8VV1ZVXtrKrTbqbMcVV1cVVdXlXvHVtNAAAA2BgO2leBqtqU5MwkT0yyK8mOqtre3VcslLlnkt9Kcnx3/1VVfdP+qjAAAACsZ8uMaB+bZGd3X9Xd1yc5J8kJK8o8K8mbuvuvkqS7PzO2mgAAALAxLBO0D01yzcL6rnnbovsn+SdVdX5VXVRVzxlVQQAAANhI9jl1PEmtsq1XOc4jk3x3krsk+WBVXdDdH7/JgapOTnJykhxxxBG3vLYAAACwzi0zor0ryeEL64cluXaVMu/s7r/v7s8meV+Sh608UHef1d1bu3vr5s2bb22dAQAAYN1aJmjvSHJ0VR1VVQcnOSnJ9hVl/iTJ46rqoKq6a5J/luSjY6sKAAAA698+p453941VdWqS85JsSnJ2d19eVafM+7d190er6p1JLkny1SSv7u7L9mfFAQAAYD1a5jva6e5zk5y7Ytu2FesvT/LycVUDAACAjWeZqeMAAADAkgRtAAAAGEjQBgAAgIEEbQAAABhI0AYAAICBBG0AAAAYSNAGAACAgQRtAAAAGEjQBgAAgIEEbQAAABhI0AYAAICBBG0AAAAYSNAGAACAgQRtAAAAGEjQBgAAgIEEbQAAABhI0AYAAICBBG0AAAAYSNAGAACAgQRtAAAAGEjQBgAAgIEEbQAAABhI0AYAAICBBG0AAAAYSNAGAACAgQRtAAAAGEjQBgAAgIEEbQAAABhI0AYAAICBBG0AAAAYSNAGAACAgZYK2lV1fFVdWVU7q+q0VfYfV1Wfr6qL58cvja8qAAAArH8H7atAVW1KcmaSJybZlWRHVW3v7itWFP2z7v6X+6GOAAAAsGEsM6J9bJKd3X1Vd1+f5JwkJ+zfagEAAMDGtEzQPjTJNQvru+ZtKz26qj5SVe+oqgcNqR0AAABsMPucOp6kVtnWK9b/IsmR3f3FqnpKkrckOfprDlR1cpKTk+SII464hVUFAACA9W+ZEe1dSQ5fWD8sybWLBbr7C939xXn53CR3rqpDVh6ou8/q7q3dvXXz5s23odoAAACwPi0TtHckObqqjqqqg5OclGT7YoGq+paqqnn52Pm4nxtdWQAAAFjv9jl1vLtvrKpTk5yXZFOSs7v78qo6Zd6/LckzkvxEVd2Y5EtJTuruldPLAQAA4A5vme9o754Ofu6KbdsWln8zyW+OrRoAAABsPMtMHQcAAACWJGgDAADAQII2AAAADCRoAwAAwECCNgAAAAwkaAMAAMBAgjYAAAAMJGgDAADAQII2AAAADCRoAwAAwECCNgAAAAwkaAMAAMBAgjYAAAAMJGgDAADAQII2AAAADCRoAwAAwECCNgAAAAwkaAMAAMBAgjYAAAAMJGgDAADAQII2AAAADCRoAwAAwECCNgAAAAwkaAMAAMBAgjYAAAAMJGgDAADAQII2AAAADCRoAwAAwECCNgAAAAwkaAMAAMBAgjYAAAAMtFTQrqrjq+rKqtpZVaftpdyjquorVfWMcVUEAACAjWOfQbuqNiU5M8mTkxyT5JlVdczNlHtZkvNGVxIAAAA2imVGtI9NsrO7r+ru65Ock+SEVcr9dJI3JvnMwPoBAADAhrJM0D40yTUL67vmbf+oqg5N8vQk2/Z2oKo6uaourKoLr7vuultaVwAAAFj3lgnatcq2XrH+X5O8qLu/srcDdfdZ3b21u7du3rx52ToCAADAhnHQEmV2JTl8Yf2wJNeuKLM1yTlVlSSHJHlKVd3Y3W8ZUksAAADYIJYJ2juSHF1VRyX5ZJKTkjxrsUB3H7V7uapek+RtQjYAAAAHon0G7e6+sapOzXQ38U1Jzu7uy6vqlHn/Xr+XDQAAAAeSZUa0093nJjl3xbZVA3Z3P/e2VwsAAAA2pmVuhgYAAAAsSdAGAACAgQRtAAAAGEjQBgAAgIEEbQAAABhI0AYAAICBBG0AAAAYSNAGAACAgQRtAAAAGEjQBgAAgIEEbQAAABhI0AYAAICBBG0AAAAYSNAGAACAgQRtAAAAGEjQBgAAgIEEbQAAABhI0AYAAICBBG0AAAAYSNAGAACAgQRtAAAAGEjQBgAAgIEOWusKwIFoy2lvX+sqsJ9cfcZT17oKAACsMSPaAAAAMJCgDQAAAAMJ2gAAADCQoA0AAAADCdoAAAAwkKANAAAAAwnaAAAAMNBSQbuqjq+qK6tqZ1Wdtsr+E6rqkqq6uKourKrHjq8qAAAArH8H7atAVW1KcmaSJybZlWRHVW3v7isWir07yfbu7qp6aJI/SvLA/VFhAAAAWM+WGdE+NsnO7r6qu69Pck6SExYLdPcXu7vn1W9I0gEAAIAD0DJB+9Ak1yys75q33URVPb2qPpbk7UmeN6Z6AAAAsLEsE7RrlW1fM2Ld3W/u7gcmeVqSX171QFUnz9/hvvC66667ZTUFAACADWCZoL0ryeEL64clufbmCnf3+5Lcr6oOWWXfWd29tbu3bt68+RZXFgAAANa7ZYL2jiRHV9VRVXVwkpOSbF8sUFXfVlU1Lz8iycFJPje6sgAAALDe7fOu4919Y1WdmuS8JJuSnN3dl1fVKfP+bUm+L8lzquqGJF9KcuLCzdEAAADggLHPoJ0k3X1uknNXbNu2sPyyJC8bWzUAAADYeJaZOg4AAAAsSdAGAACAgQRtAAAAGEjQBgAAgIEEbQAAABhI0AYAAICBBG0AAAAYSNAGAACAgQRtAAAAGEjQBgAAgIEEbQAAABhI0AYAAICBBG0AAAAYSNAGAACAgQRtAAAAGEjQBgAAgIEEbQAAABhI0AYAAICBBG0AAAAYSNAGAACAgQRtAAAAGEjQBgAAgIEEbQAAABhI0AYAAICBBG0AAAAYSNAGAACAgQRtAAAAGEjQBgAAgIEEbQAAABhI0AYAAICBBG0AAAAYaKmgXVXHV9WVVbWzqk5bZf+zq+qS+fGBqnrY+KoCAADA+rfPoF1Vm5KcmeTJSY5J8syqOmZFsU8k+c7ufmiSX05y1uiKAgAAwEZw0BJljk2ys7uvSpKqOifJCUmu2F2guz+wUP6CJIeNrCQAABvXltPevtZVYD+5+oynrnUVYF1aZur4oUmuWVjfNW+7OT+S5B23pVIAAACwUS0zol2rbOtVC1Y9PlPQfuzN7D85yclJcsQRRyxZRQAAANg4lhnR3pXk8IX1w5Jcu7JQVT00yauTnNDdn1vtQN19Vndv7e6tmzdvvjX1BQAAgHVtmaC9I8nRVXVUVR2c5KQk2xcLVNURSd6U5Ae7++PjqwkAAAAbwz6njnf3jVV1apLzkmxKcnZ3X15Vp8z7tyX5pST3TvJbVZUkN3b31v1XbQAAAFiflvmOdrr73CTnrti2bWH5R5P86NiqAQAAwMazzNRxAAAAYEmCNgAAAAwkaAMAAMBAgjYAAAAMJGgDAADAQII2AAAADCRoAwAAwECCNgAAAAwkaAMAAMBAgjYAAAAMJGgDAADAQII2AAAADCRoAwAAwECCNgAAAAwkaAMAAMBAgjYAAAAMJGgDAADAQII2AAAADCRoAwAAwECCNgAAAAwkaAMAAMBAgjYAAAAMJGgDAADAQII2AAAADCRoAwAAwECCNgAAAAwkaAMAAMBAgjYAAAAMJGgDAADAQII2AAAADLRU0K6q46vqyqraWVWnrbL/gVX1war6clW9cHw1AQAAYGM4aF8FqmpTkjOTPDHJriQ7qmp7d1+xUOxvkjw/ydP2Sy0BAABgg1hmRPvYJDu7+6ruvj7JOUlOWCzQ3Z/p7h1JbtgPdQQAAIANY5mgfWiSaxbWd83bAAAAgBWWCdq1yra+NSerqpOr6sKquvC66667NYcAAACAdW2ZoL0ryeEL64clufbWnKy7z+rurd29dfPmzbfmEAAAALCuLRO0dyQ5uqqOqqqDk5yUZPv+rRYAAABsTPu863h331hVpyY5L8mmJGd39+VVdcq8f1tVfUuSC5N8Y5KvVtULkhzT3V/Yj3UHAACAdWefQTtJuvvcJOeu2LZtYfmvM00pBwAAgAPaMlPHAQAAgCUtNaINAACwXmw57e1rXQX2k6vPeOpaV2EII9oAAAAwkKANAAAAAwnaAAAAMJCgDQAAAAMJ2gAAADCQoA0AAAADCdoAAAAwkKANAAAAAwnaAAAAMJCgDQAAAAMJ2gAAADCQoA0AAAADCdoAAAAwkKANAAAAAwnaAAAAMJCgDQAAAAMJ2gAAADCQoA0AAAADCdoAAAAwkKANAAAAAwnaAAAAMJCgDQAAAAMJ2gAAADCQoA0AAAADCdoAAAAwkKANAAAAAwnaAAAAMJCgDQAAAAMdtNYVAOC223La29e6CuwnV5/x1LWuAgBwCy01ol1Vx1fVlVW1s6pOW2V/VdVvzPsvqapHjK8qAAAArH/7DNpVtSnJmUmenOSYJM+sqmNWFHtykqPnx8lJXjW4ngAAALAhLDOifWySnd19VXdfn+ScJCesKHNCkt/ryQVJ7llV9xlcVwAAAFj3lgnahya5ZmF917ztlpYBAACAO7xlboZWq2zrW1EmVXVypqnlSfLFqrpyifNz+zskyWfXuhK3h3rZWtfggHDAtKdEm7qdaFOMdsC0Ke3pdqNNMdIB056SDdemjry5HcsE7V1JDl9YPyzJtbeiTLr7rCRnLXFO1lBVXdjdW9e6HtwxaE+Mpk0xmjbFaNoUI2lPG9MyU8d3JDm6qo6qqoOTnJRk+4oy25M8Z777+Hck+Xx3f2pwXQEAAGDd2+eIdnffWFWnJjkvyaYkZ3f35VV1yrx/W5Jzkzwlyc4k/5Dkh/dflQEAAGD9WmbqeLr73ExhenHbtoXlTvJTY6vGGjK9n5G0J0bTphhNm2I0bYqRtKcNqKaMDAAAAIywzHe0AQAAgCUJ2ndAVfXF+ed9q+oNeyl3z6r6yduvZmwEVfX8qvpoVb1+Px1/a1X9xrx8XFX98/1xHtibqnpBVd11resBANwxCdp3YN19bXc/Yy9F7plE0Galn0zylO5+9ugDV9VB3X1hdz9/3nRcEkGbtfCCJIL2OlVVV1fVIVW1paouux3Pe35VbZ2Xz62qe+6lrM6adWgN287jquryqrq4qu6yH8/zgfnnlqp61v46z4FsrdrQRlJVT6uqY9a6HuudoH0HtniBqKoHVdWH5jeAS6rq6CRnJLnfvO3lc7mfr6odc5mXLBzno1X1O/ObyLt2v4nMo59XzOXPWavflTGqaluSb02yvapeVFUfqKoPzz8fMJf586p60MJzzq+qR1bVvarqLXNbuKCqHjrvP72qzqqqdyX5vXkU+21VtSXJKUl+dm6Dj6uqzVX1xrkN7qiqx8zH+M65zMVzfe5+O//TsKSqes7cBj5SVb9fVUdW1bvnbe+uqiPmcq+pqldV1Xuq6qr5NT57vta8ZuF4X6yql1XVRVX1p1V17Nzmrqqq75nLbKqqly9cu3583n7cXPYNVfWxqnp9TZ6f5L5J3jOff9Ncn8uq6tKq+tk1+Kdjnenup3T3/91LEZ01LHp2kl/t7od395dGH7yqNiVJd+/unN6SRNBmrTwtiaC9L93tcQd7JPni/HNLksvm5Vcmefa8fHCSuyzun7c/KdNdDStTJ8zbkvyLudyNSR4+l/ujJD8wL1+b5Ovm5Xuu9e/uMaT9XJ3kkCTfmOSgedsTkrxxXv7ZJC+Zl++T5OMLbew/zMvfleTiefn0JBclucu8flySty3se+HCuf8gyWPn5SOSfHRefmuSx8zLd9tdL4/19UjyoCRXJjlkXr/X/Nr90Lz+vCRvmZdfk+Sc+XpzQpIvJHnIfO25aOF600mePC+/Ocm7ktw5ycMW2tjJSV48L39dkguTHDW3tc8nOWw+7gcX2tfVC/V8ZJL/sfB7uJbdvu3mLfNrfnmSkxdfn/n952NJXpvkkiRvSHLXVV7DrUnOn5dPn8u/ay7zvUl+JcmlSd6Z5M57qcv5SbauqMM3JHl7ko8kuSzJiUmen+T6+Zjvmcs/aW5jf5Hkj5PcbeE4L5m3X5rkgfP270xy8fz4cJK7r/VrsdEe66XtJPnRJH+T5BNJXp/pferdC6/5CXO5lyX5yYXnnZ7k32S6Dr58bl+XJjlx3n9ckvdkem+8Yt62+zPeBZmubxdnel/eNB9jx/z7/vhc7j5J3jeXuyzJ49b6dVtPj/XShubnPirJBzJdaz6U5O5Jvj7Jf5uf/+Ekj5/LPneu+1vndndqkp+by1yQ5F5zufOTvGJuAx+dz/GmJP87yX9cOPcPzOe8OMlvJ9m0u70l+U9znS5I8s2ZZiLubu8XJ7lfpmviFfO/0zlr/bqul4cR7QPHB5P8YlW9KMmRvXpv65Pmx4czvTk8MMnR875PdPfF8/JFmS4+yfQf6vVV9QOZwjh3HPdI8sfzrIhXZApRydTR8q/n5e/P9IEySR6b5PeTpLv/Z5J7V9U95n3bb6bNrfSEJL9ZVRcn2Z7kG+fR6/cn+bV5JPKe3a2trU/fleQN3f3ZJOnuv0ny6EwfEpOpfTx2ofxbe3onvzTJp7v70u7+aqYPPFvmMtdn+nCSudx7u/uGeXl3mSclec7cbv48yb2z59r1oe7eNR/34oXnLLoqybdW1Sur6vhMoZ/bz/O6+5GZPqw+v6ruvWL/A5Kc1d0PzfTaLPOVp/sleWqmTpzXZQrDD0nypXn7LXF8kmu7+2Hd/eAk7+zu38jU0fz47n58VR2S5MVJntDdj8jU2fNzC8f47Lz9VUleOG97YZKf6u6HJ3ncXDdumXXRdrr71Znes36+p69d/b8kT59f88cn+S9VVZk6F09ceOru99DvTfLwTB2IT0jy8qq6z1zm2CT/rrtXjh6eluTPehpBf0WSH0ny+e5+VKYw9WNVdVSmUe/z5nb2sEzXQfZYF22oqg5O8t+T/Ex3724HX8r855Pn5z8zyWur6uvnpz040+t7bKYw/A/d/e2ZPvM/Z+Hw13f3v0iyLcmfzMd8cJLnVtW9q+qfZmqXj5nbyVcyzdBIpo7GC+Y6vS/Jj3X3B7KnvT+8u/9Ppvb47fO/0ylL/BsdEATtA0R3/0GS78n0n/a8qvquVYpVkpfO/2ke3t3f1t2/O+/78kK5r2TP32B/apIzM40IXVRVS/1tdjaEX8705vDgJP8qU69quvuTST43Tw0/MdMHh2RqPyvt/vuBf7/kOe+U5NELbfDQ7v677j4j04jBXZJcUFUPvHW/EvtZZc9rfnMW9+++rnw1N73GfDV7rjE3zGH8JuXm4Ly7TCX56YV2c1R3v2vFOZKbXrv2VKj7bzN9AD0/0weQV+/jd2Cs51fV7tGSw7Onk2S3a7r7/fPy63LTzpqb846FDplNuWlnzZZbWL9Lkzxh/grD47r786uU+Y5M0yjfP3f4/FCSIxf2v2n+udhRrQPxtluvbaeS/OequiTJnyY5NMk3d/eHk3xTTTerfViSv+3uv5rr9Yfd/ZXu/nSS92YKy8nUWfiJJc55cx2OO5L8cFWdnuQh3f13S/4OB4r10oYekORT3b0jSbr7C/M1YXEQ42NJ/jLJ/efnvGf+jHRdptkNb72Z82xf2H55d3+qu7+cqZP58CTfnelz/I65/Xx3pq8RJlNn99vm5cXr10oG3lYhaB8gqupbk1w198JvT/LQJH+XaVrKbucleV5V3W1+zqFV9U17Oeadkhze3e9J8guZbq52t/30K3D7u0eST87Lz12x75xMr/k9uvvSedv7MveAVtVxmUZw9jUyuLINvivT9KfMx3n4/PN+82jnyzKNFAna69O7k3z/7hGBqrpXpmlwJ837n53kf+2H856X5Ceq6s7zee9fVd+wj+f8Y9ubRyPv1N1vTPLvkzxiP9SRVczXiidk6mB7WKYZVV+/otjKzpvd6zdmz+eYlc9Z7JBZ2VlzizqEu/vjmT6EXprkpVX1S6v9Kpm+frC7s+eY7v6RlfXJQmePDsTbZp23nWcn2ZzkkfMI4acXzvOGJM/Ivjuqd1u2o3rVDsfufl+mrwF+MsnvV9Vz9n6YA8c6a0M311G9t7axsoN6sfP6oFXK3VyndiV57ULbeUB3nz6XWaz/qp3VMwNvqxC0DxwnJrls7ql6YJLf6+7PZep9v6yqXj6PAP1Bkg9W1aWZ3gz2dtOpTUleN5f9cJJX9N5vHMPG8iuZPlS+P9NrvegNmcLTHy1sOz3J1rkH/4xMIzr78tYkT59vcva4TN/x2Trf0OqK7Jl+9IK5nX4k06yMd9zaX4r9p7svzzR97b3za/VrmV7TH57bxQ8m+Zn9cOpXZ/pu2F/MX3X47ez7A/FZSd5RVe/JNNp0/nx9fE2Sf7sf6sjq7pFpVO8f5qD5HauUOaKqHj0vPzN7OmuuzvShLkm+b39VsKrum2lK5uuS/Gr2dMQsdhRekOQxVfVt83PuWlX3/5qD3fS4OhBvm/Xcdu6R5DPdfUNVPT43nd1wTqb3z2dkei9Npo7qE+cbM27OFIw/tI9zrDZY8jUdjlV15FyX30nyu9GRuGg9taGPJblvVT0qSarq7nNYXRzEuH+m+9dcOeB8i96d5Bm7B9dqurntkft4zmJntYG3m6G34Q6ou+82/7w603cw0t0vTfLSVco+a8X6ryf59VUO++CFMr+6sH2ZKTRsIN29ZV78bPZMT0qmkb7dZT6dFdePnr6Pe8Iqxzt9xfr5mabo7h4peuiKp5y4Yj3d/dPL1Z611t2vzXQjmEVf81WV7n7uwvLVuek1ZnHf3RaWT19xjN3Xuq8m+cX5sej8+bG7/KkLy6/MdAO/3Xz4XBvvTHLK3BFzZabAutJHk/xQVf12phv4vGre/pIkv1tVv5hpquz+8pBM35n9apIbkvzEvH13Z82nevqe9nOT/GFVfd28/8VJPr6X475gDmFfydRRpAPxllnPbef1Sd5aVRdm+k70x3bv6O7L53uPfLK7PzVvfnOm+1l8JNOo5i9091/vY5bDJUlunDs1X5Pps9uWTB2OleS6THeGPi7Jz1fVDZlubGVEe49104a6+/qqOjHJK2v6yz5fyjTa/ltJts2DWjcmeW53f3l6icfo7iuq6sVJ3jWH5hsyfY3qL/fytHOS/M781ZeTMv1b3CPT6LiBt1ntmQ0AAAAA3FamjgMAAMBApo4DAAeUqnpzpr+1vuhF3X3eWtSHjUPb4bbShg4cpo4DAADAQKaOAwAAwECCNgAAAAwkaAMAAMBAgjYAAAAMJGgDAADAQP8felIb3Y0F+wQAAAAASUVORK5CYII=\n",
"text/plain": [
"
"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"# display the weight of each component in the projection\n",
"plt.bar(X.columns\n",
" , (pca.components_)[0]\n",
" )\n",
"plt.show()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The above plot represents the weights of the different features (or their associated coefficients) in the projection the 6 dimensions on a 1D axis representing the popularity score."
]
},
{
"cell_type": "code",
"execution_count": 15,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 355
},
"id": "uQvLji7i9cMw",
"outputId": "1a6210c6-5f3c-4251-aa54-ec152a35acd2"
},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"/home/mbien/.local/lib/python3.8/site-packages/scipy/stats/morestats.py:1676: UserWarning: p-value may not be accurate for N > 5000.\n",
" warnings.warn(\"p-value may not be accurate for N > 5000.\")\n"
]
},
{
"data": {
"image/png": "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\n",
"text/plain": [
"
"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"# plot the distribution of the 1D-popularity projection, and showcase its Shapiro-Wilk score\n",
"W, p_val = shapiro(pop_df.pop_score)\n",
"\n",
"plt.hist(pop_df.pop_score, bins=100, label=r\"$W_{Shapiro-Wilk}\"+\"={0:.3f}$\".format(W))\n",
"plt.legend(fontsize=12)\n",
"plt.show()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The popularity scores obtained after the 1D PCA projection follow a gaussian distribution (as underlined by the Shapiro-Wilk score close to 1). This measure is then $z$-score normalized to make it follow a normal law $\\mathcal{N}(0,1)$."
]
},
{
"cell_type": "code",
"execution_count": 16,
"metadata": {
"id": "douDo7u89cMx"
},
"outputs": [],
"source": [
"# as the distribution is ~ normal (shapiro close to 1), standardize the score (z-score normalization)\n",
"pop_df[\"pop_score\"] = (pop_df[\"pop_score\"]-np.mean(pop_df[\"pop_score\"]))/(np.std(pop_df[\"pop_score\"]))"
]
},
{
"cell_type": "code",
"execution_count": 17,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 175
},
"id": "cEMHUN699cMy",
"outputId": "720c2084-cdcd-4182-fcd2-fcf249fd9359"
},
"outputs": [
{
"data": {
"text/html": [
"
\n",
"\n",
"
\n",
" \n",
"
\n",
"
\n",
"
date_created
\n",
"
genre_top
\n",
"
listens
\n",
"
favorites
\n",
"
comments
\n",
"
album_listens
\n",
"
album_favorites
\n",
"
album_comments
\n",
"
pop_score
\n",
"
\n",
"
\n",
"
track_id
\n",
"
\n",
"
\n",
"
\n",
"
\n",
"
\n",
"
\n",
"
\n",
"
\n",
"
\n",
"
\n",
" \n",
" \n",
"
\n",
"
125532
\n",
"
2015-09-25 16:18:52
\n",
"
Electronic
\n",
"
1352
\n",
"
0
\n",
"
0
\n",
"
4473
\n",
"
0
\n",
"
0
\n",
"
-0.275386
\n",
"
\n",
"
\n",
"
85157
\n",
"
2013-05-30 15:50:26
\n",
"
Experimental
\n",
"
230
\n",
"
2
\n",
"
0
\n",
"
30149
\n",
"
4
\n",
"
11
\n",
"
0.213893
\n",
"
\n",
"
\n",
"
44980
\n",
"
2011-03-04 13:55:33
\n",
"
Experimental
\n",
"
102
\n",
"
1
\n",
"
0
\n",
"
1292
\n",
"
0
\n",
"
0
\n",
"
-1.407466
\n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" date_created genre_top listens favorites comments \\\n",
"track_id \n",
"125532 2015-09-25 16:18:52 Electronic 1352 0 0 \n",
"85157 2013-05-30 15:50:26 Experimental 230 2 0 \n",
"44980 2011-03-04 13:55:33 Experimental 102 1 0 \n",
"\n",
" album_listens album_favorites album_comments pop_score \n",
"track_id \n",
"125532 4473 0 0 -0.275386 \n",
"85157 30149 4 11 0.213893 \n",
"44980 1292 0 0 -1.407466 "
]
},
"execution_count": 17,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"pop_df.sample(3)"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "JyQB9_3CFtjV"
},
"source": [
"## Generate samples for finetuning"
]
},
{
"cell_type": "code",
"execution_count": 18,
"metadata": {
"id": "D2i668J6FlSV"
},
"outputs": [],
"source": [
"from sklearn.model_selection import train_test_split\n",
"\n",
"train, test = train_test_split(pop_df.pop_score, test_size=0.05)\n",
"train.to_csv(\"train.csv\")\n",
"test.to_csv(\"test.csv\")"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "AzPb4uWN9cMz"
},
"source": [
"## Feature selection for regressions "
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "Z_uvpeKG9cM0"
},
"source": [
"Now we are going to proceed to the regression analysis. The features of our dataset are at a number of 518. There are : \n",
"- *chroma_cens* (max, min, mean, median, skew, kurtosis, and standard deviation) (N = 1...12): corresponds to the chroma energy normalized, how much energy of pitch class N is present in the signal\n",
"- *Mel-frequency cepstral coefficients (MFCCs)* (max, min, mean, median, skew, kurtosis, and standard deviation) (N = 1....20) : N-th coefficient that describe the short-term power spectrum of a sound, that itself describes how power of a signal or time series is distributed over frequency\n",
"- *Root-mean-square (RMS) energy* (max, min, mean, median, skew, kurtosis, and standard deviation) (N=1...20) : corresponds to how 'loud' the music is\n",
"- *Spectral Bandwidth* (max, min, mean, median, skew, kurtosis, and standard deviation) : band width at one-half the peak maximum\n",
"- *Spectral Centroid* (max, min, mean, median, skew, kurtosis, and standard deviation) : an acoustical descriptor of timbre. Estimates the center of mass of the spectrum (in Hz)\n",
"- *Spectral Contrast* (max, min, mean, median, skew, kurtosis, and standard deviation) : works on octaves and considers spectral peak, spectral valley and their difference in each sub-band\n",
"- *Tonnetz* (max, min, mean, median, skew, kurtosis, and standard deviation) (N=0...5) : N represents : fifth x-axis, fifth y-axis, minor x-axis, minor y-axis, major x-axis, major y-axis\n",
"- *Zero Crossing Rate* (max, min, mean, median, skew, kurtosis, and standard deviation) : rate at which a signal changes from positive to zero to negative or from negative to zero to positive\n",
"\n",
"As it is computationnaly too expensive to compute a regression with all the features, we are need to reduce the dimensionality of each data point. To do that, we observe that some features are potentially strongly correlated, and including both of them in the analysis would be unnecessary. For example, the mean ZCR and median ZCR are likely to be somehow linked. Our assumption is that we can find a smaller set of features that describe well enough our dataset.\n",
"\n",
"To find the features, we will compute the Pearson's correlation between each feature, pairwise. It gives us information about the joint variablitiy of two random variables, showing linear relationships between them. For two random variables X and Y, the correlation is computed as follow :\n",
"\n",
"$$corr(X,Y) = \\frac{cov(X, Y)}{std(X) * std(Y)} $$\n",
"\n",
"Where the covariance is defined as : $$cov(X,Y) = \\frac{\\sum(x_i - \\overline{x})(y_i - \\overline{y})}{N}$$\n",
"\n",
"A first limitation of this approach is that we only account for linear relationships, not more complex ones.\n"
]
},
{
"cell_type": "code",
"execution_count": 18,
"metadata": {
"colab": {
"base_uri": "https://localhost:8080/",
"height": 287
},
"id": "FASXF2LW9cM0",
"outputId": "38fe4cb6-aab3-42f4-c3b9-53f1bd7c4e13"
},
"outputs": [
{
"data": {
"text/html": [
"
"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"fig, ax = plt.subplots(figsize=(10,10))\n",
"im = ax.imshow(corr_matrix.abs())\n",
"im.set_clim(0, 1)\n",
"ax.grid(False)\n",
"cbar = ax.figure.colorbar(im, ax=ax, format='% .2f')\n",
"plt.show()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"That shows that some features are redundant, as not only the diagional has a green color.\n",
"\n",
"Now, to select a meaningful set of features, that describes the data well enough, we are taking the following approach, greatly inspired by [this](https://towardsdatascience.com/escape-the-correlation-matrix-into-feature-space-4d71c51f25e5). They summarize the approach this way : \n",
"1. Take the absolute value of our correlation matrix, and subtract each value from 1. It is handily transformed into a distance matrix!\n",
"2. We can then use PCA to reduce our NxN matrix to Nx2.\n",
"3. Plot each feature’s location using the two principal components.\n",
"4. Use Feature Agglomeration to generate feature clusters.\n",
"5. Color each feature by its cluster.\n",
"6. Draw lines to represent relationships "
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "EE89sgyX9cM3"
},
"source": [
"We re-use the code from [here](https://gist.github.com/MattJBritton/b6944218903312f6220bfb48706b593c#file-minimal_example-py) for the nice visualisations. The code is in a separate notebook (milestone3_feature_clustering.ipynb) to avoid overload this one. We directly use here the appropriate features."
]
},
{
"cell_type": "code",
"execution_count": 21,
"metadata": {
"id": "q012UZED9cM6"
},
"outputs": [],
"source": [
"selected_features = ['chroma_stft_mean_01', 'chroma_stft_std_01', 'chroma_cqt_kurtosis_03', 'chroma_stft_max_01', 'chroma_stft_kurtosis_01', 'chroma_stft_kurtosis_03', 'chroma_cqt_mean_01', 'chroma_cqt_mean_07','chroma_cqt_kurtosis_01', 'chroma_cqt_kurtosis_08']\n",
"selected_features = selected_features + ['mfcc_max_01', 'mfcc_max_05', 'mfcc_kurtosis_03', 'mfcc_min_04', 'mfcc_skew_09', 'mfcc_max_02', 'mfcc_max_03', 'mfcc_skew_04', 'mfcc_skew_03', 'mfcc_kurtosis_01']\n",
"selected_features = selected_features + ['rmse_max_01', 'rmse_kurtosis_01', 'rmse_min_01']\n",
"selected_features = selected_features + ['spectral_contrast_max_01', 'spectral_contrast_max_07', 'spectral_contrast_kurtosis_01', 'spectral_contrast_min_01', 'spectral_rolloff_kurtosis_01', 'spectral_contrast_max_06', 'spectral_rolloff_max_01', 'spectral_contrast_kurtosis_06', 'spectral_contrast_kurtosis_07', 'spectral_contrast_skew_07']\n",
"selected_features = selected_features + ['tonnetz_max_01', 'tonnetz_kurtosis_01', 'tonnetz_mean_02', 'tonnetz_min_01', 'tonnetz_mean_01', 'tonnetz_mean_03', 'tonnetz_skew_03', 'tonnetz_skew_04', 'tonnetz_skew_05', 'tonnetz_skew_06']"
]
},
{
"cell_type": "code",
"execution_count": 26,
"metadata": {},
"outputs": [],
"source": [
"features_flatten = features.copy()\n",
"features_flatten.columns = ['_'.join(col) for col in features.columns.values]"
]
},
{
"cell_type": "code",
"execution_count": 36,
"metadata": {},
"outputs": [],
"source": [
"chroma = features[['chroma_stft', 'chroma_cqt', 'chroma_cens']]\n",
"chroma.columns = ['_'.join(col) for col in chroma.columns.values]\n",
"\n",
"mfcc = features['mfcc']\n",
"mfcc.columns = ['_'.join(col) for col in mfcc.columns.values]\n",
"\n",
"rmse = features['rmse']\n",
"rmse.columns = ['_'.join(col) for col in rmse.columns.values]\n",
"\n",
"spectral = features[['spectral_contrast', 'spectral_rolloff', 'spectral_centroid', 'spectral_bandwidth']]\n",
"spectral.columns = ['_'.join(col) for col in spectral.columns.values]\n",
"\n",
"tonnetz = features['tonnetz']\n",
"tonnetz.columns = ['_'.join(col) for col in tonnetz.columns.values]\n",
"\n",
"zcr = features['zcr']\n",
"zcr.columns = ['_'.join(col) for col in zcr.columns.values]"
]
},
{
"cell_type": "code",
"execution_count": 37,
"metadata": {
"id": "mMJXkSQE9cM7"
},
"outputs": [],
"source": [
"# Normalizing selected features for regression\n",
"df_selected_features = features_flatten[selected_features]\n",
"x = df_selected_features.values \n",
"min_max_scaler = preprocessing.MinMaxScaler()\n",
"x_scaled = min_max_scaler.fit_transform(x)\n",
"df_selected_features_normalized = pd.DataFrame(x_scaled)"
]
},
{
"cell_type": "code",
"execution_count": 38,
"metadata": {
"id": "VMfctIis9cM7"
},
"outputs": [],
"source": [
"df_selected_features_normalized.index = tracks[\"track\"]['genre_top'].index\n",
"df_selected_features_normalized = df_selected_features_normalized[df_selected_features_normalized.index.isin(pop_df[\"pop_score\"].index)]\n",
"X = df_selected_features_normalized\n",
"y = pop_df[\"pop_score\"]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Regressions "
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We decide to use regression analysis to explore as a statistical method that allows us to examine the relationship between the variables of interest : popularity and musical features. Our dependent variable is the popularity score as computed previously, whereas the independent variables are all the features selected by the previous clustering."
]
},
{
"cell_type": "code",
"execution_count": 39,
"metadata": {
"id": "QfdxRMCT9cM8",
"outputId": "f176dabc-9bde-4daa-c64e-5474de698190",
"scrolled": true
},
"outputs": [
{
"data": {
"text/html": [
"
"
],
"text/plain": [
" chroma_cens_kurtosis_01 chroma_cens_kurtosis_02 \\\n",
"track_id \n",
"2 0.001033 0.002501 \n",
"3 0.000435 0.000941 \n",
"5 0.000281 0.000649 \n",
"10 0.000640 0.000574 \n",
"20 0.000199 0.000606 \n",
"\n",
" chroma_cens_kurtosis_03 chroma_cens_kurtosis_04 \\\n",
"track_id \n",
"2 0.001246 0.001021 \n",
"3 0.001300 0.001243 \n",
"5 0.000948 0.000865 \n",
"10 0.002340 0.000649 \n",
"20 0.001219 0.000765 \n",
"\n",
" chroma_cens_kurtosis_05 chroma_cens_kurtosis_06 \\\n",
"track_id \n",
"2 0.001067 0.000873 \n",
"3 0.001121 0.000708 \n",
"5 0.001210 0.000997 \n",
"10 0.001031 0.001039 \n",
"20 0.000845 0.000714 \n",
"\n",
" chroma_cens_kurtosis_07 chroma_cens_kurtosis_08 \\\n",
"track_id \n",
"2 0.000798 0.001470 \n",
"3 0.000771 0.001047 \n",
"5 0.000236 0.000535 \n",
"10 0.000865 0.001189 \n",
"20 0.000384 0.000564 \n",
"\n",
" chroma_cens_kurtosis_09 chroma_cens_kurtosis_10 ... \\\n",
"track_id ... \n",
"2 0.003169 0.002444 ... \n",
"3 0.002547 0.002038 ... \n",
"5 0.003005 0.002274 ... \n",
"10 0.002886 0.002806 ... \n",
"20 0.002495 0.001236 ... \n",
"\n",
" tonnetz_std_04 tonnetz_std_05 tonnetz_std_06 zcr_kurtosis_01 \\\n",
"track_id \n",
"2 0.148243 0.130829 0.101443 0.000915 \n",
"3 0.177944 0.152235 0.150073 0.000568 \n",
"5 0.107253 0.135841 0.124322 0.001040 \n",
"10 0.210157 0.192545 0.117085 0.002773 \n",
"20 0.273332 0.241482 0.181304 0.002208 \n",
"\n",
" zcr_max_01 zcr_mean_01 zcr_median_01 zcr_min_01 zcr_skew_01 \\\n",
"track_id \n",
"2 0.454097 0.097468 0.079564 0.000000 0.189929 \n",
"3 0.461007 0.096271 0.071390 0.000000 0.185952 \n",
"5 0.368707 0.060433 0.046322 0.000000 0.191032 \n",
"10 0.446693 0.088225 0.080109 0.000000 0.205412 \n",
"20 0.464462 0.053726 0.044687 0.003724 0.201654 \n",
"\n",
" zcr_std_01 \n",
"track_id \n",
"2 0.143937 \n",
"3 0.162532 \n",
"5 0.104805 \n",
"10 0.095226 \n",
"20 0.072089 \n",
"\n",
"[5 rows x 518 columns]"
]
},
"execution_count": 36,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"min_max_scaler_q = preprocessing.MinMaxScaler()\n",
"feature_flatten_normalized = min_max_scaler_q.fit_transform(features_flatten)\n",
"feature_flatten_normalized = pd.DataFrame(feature_flatten_normalized)\n",
"feature_flatten_normalized.columns = features_flatten.columns\n",
"feature_flatten_normalized.index = features_flatten.index\n",
"feature_flatten_normalized.head()"
]
},
{
"cell_type": "code",
"execution_count": 37,
"metadata": {},
"outputs": [],
"source": [
"features_genre_relevant = feature_flatten_normalized[tracks[\"track\"]['genre_top'] == 'Pop'][df_merged.index[df_merged.index != 'const']]\n",
"mult = features_genre_relevant * df_merged['Pop'] \n",
"pop_score_pred = mult.sum(axis=1) "
]
},
{
"cell_type": "code",
"execution_count": 38,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"track_id\n",
"10 -5.404185\n",
"213 -5.742983\n",
"821 -5.629483\n",
"822 -5.652588\n",
"823 -5.532414\n",
" ... \n",
"154410 -5.358343\n",
"154411 -5.563459\n",
"154412 -5.206557\n",
"154413 -5.894569\n",
"154414 -6.035722\n",
"Length: 2321, dtype: float64"
]
},
"execution_count": 38,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"pop_score_pred[~no_album][tracks[\"track\"]['genre_top'] == 'Pop']"
]
},
{
"cell_type": "code",
"execution_count": 39,
"metadata": {
"id": "nqp66gkA9cNP"
},
"outputs": [],
"source": [
"def n_samples_genre(n, genre, high_score_threshold, low_score_threshold):\n",
" print('Genre : ', genre)\n",
" features_genre_relevant = feature_flatten_normalized[tracks[\"track\"]['genre_top'] == genre][df_merged.index[df_merged.index != 'const']]\n",
" mult = features_genre_relevant * df_merged[genre] \n",
" pop_score_pred = mult.sum(axis=1) \n",
" \n",
" # The ones that are popular/unpopular based on the dataset\n",
" pop_df_dataset = pop_df[\"pop_score\"][tracks[\"track\"]['genre_top'] == genre][pop_df[\"pop_score\"][tracks[\"track\"]['genre_top'] == genre] > 0]\n",
" unpop_df_dataset = pop_df[\"pop_score\"][tracks[\"track\"]['genre_top'] == genre][pop_df[\"pop_score\"][tracks[\"track\"]['genre_top'] == genre] < 0]\n",
" \n",
" # The ones that are popular/unpopular based on the regression\n",
" genre_df_pred = pop_score_pred[~no_album][tracks[\"track\"]['genre_top'] == genre]\n",
" low, high = genre_df_pred.quantile([0.25,0.75])\n",
" pop_df_pred = genre_df_pred.loc[(genre_df_pred > high)]\n",
" unpop_df_pred = genre_df_pred.loc[(genre_df_pred < low)]\n",
" \n",
" # True Positive\n",
" tp = pop_df_dataset.index.intersection(pop_df_pred.index)\n",
" tp = pop_df[\"pop_score\"][tp].nlargest(1)\n",
" print('TP')\n",
" print(tp)\n",
" \n",
" # False Positive\n",
" fp = unpop_df_dataset.index.intersection(pop_df_pred.index)\n",
" fp = pop_df[\"pop_score\"][fp].nsmallest(1)\n",
" print('FP')\n",
" print(fp)\n",
" \n",
" # True Negative\n",
" tn = unpop_df_dataset.index.intersection(unpop_df_pred.index)\n",
" tn = pop_df[\"pop_score\"][tn].nsmallest(1)\n",
" print('TN')\n",
" print(tn)\n",
" \n",
" # False Negative\n",
" fn = pop_df_dataset.index.intersection(unpop_df_pred.index)\n",
" fn = pop_df[\"pop_score\"][fn].nlargest(1)\n",
" print('FN')\n",
" print(fn)\n",
" print(\"===============\")\n",
" \n",
" return [tp, fp, tn, fn]"
]
},
{
"cell_type": "code",
"execution_count": 40,
"metadata": {
"id": "5Q-eFq3n9cNR",
"outputId": "e8b136da-c589-414b-8b13-b386dfaf2671"
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Genre : Electronic\n",
"TP\n",
"track_id\n",
"69170 4.443784\n",
"Name: pop_score, dtype: float64\n",
"FP\n",
"track_id\n",
"82185 -2.565108\n",
"Name: pop_score, dtype: float64\n",
"TN\n",
"track_id\n",
"1432 -2.92627\n",
"Name: pop_score, dtype: float64\n",
"FN\n",
"track_id\n",
"28553 3.518097\n",
"Name: pop_score, dtype: float64\n",
"===============\n",
"Genre : Experimental\n",
"TP\n",
"track_id\n",
"7391 2.683178\n",
"Name: pop_score, dtype: float64\n",
"FP\n",
"track_id\n",
"155010 -2.841336\n",
"Name: pop_score, dtype: float64\n",
"TN\n",
"track_id\n",
"42956 -3.034765\n",
"Name: pop_score, dtype: float64\n",
"FN\n",
"track_id\n",
"92296 2.858244\n",
"Name: pop_score, dtype: float64\n",
"===============\n",
"Genre : Folk\n",
"TP\n",
"track_id\n",
"50952 3.00961\n",
"Name: pop_score, dtype: float64\n",
"FP\n",
"track_id\n",
"154623 -2.002789\n",
"Name: pop_score, dtype: float64\n",
"TN\n",
"track_id\n",
"38439 -2.500617\n",
"Name: pop_score, dtype: float64\n",
"FN\n",
"track_id\n",
"88863 1.738048\n",
"Name: pop_score, dtype: float64\n",
"===============\n",
"Genre : Hip-Hop\n",
"TP\n",
"track_id\n",
"24425 4.373379\n",
"Name: pop_score, dtype: float64\n",
"FP\n",
"track_id\n",
"53996 -1.684636\n",
"Name: pop_score, dtype: float64\n",
"TN\n",
"track_id\n",
"53298 -2.298028\n",
"Name: pop_score, dtype: float64\n",
"FN\n",
"track_id\n",
"10699 2.444624\n",
"Name: pop_score, dtype: float64\n",
"===============\n",
"Genre : Instrumental\n",
"TP\n",
"track_id\n",
"61053 3.348277\n",
"Name: pop_score, dtype: float64\n",
"FP\n",
"track_id\n",
"28819 -2.115002\n",
"Name: pop_score, dtype: float64\n",
"TN\n",
"track_id\n",
"24360 -2.736748\n",
"Name: pop_score, dtype: float64\n",
"FN\n",
"track_id\n",
"37111 3.091666\n",
"Name: pop_score, dtype: float64\n",
"===============\n",
"Genre : International\n",
"TP\n",
"track_id\n",
"51006 3.49511\n",
"Name: pop_score, dtype: float64\n",
"FP\n",
"track_id\n",
"149993 -2.126809\n",
"Name: pop_score, dtype: float64\n",
"TN\n",
"track_id\n",
"153476 -2.344313\n",
"Name: pop_score, dtype: float64\n",
"FN\n",
"track_id\n",
"73760 3.088421\n",
"Name: pop_score, dtype: float64\n",
"===============\n",
"Genre : Pop\n",
"TP\n",
"track_id\n",
"62436 2.997317\n",
"Name: pop_score, dtype: float64\n",
"FP\n",
"track_id\n",
"89489 -2.112164\n",
"Name: pop_score, dtype: float64\n",
"TN\n",
"track_id\n",
"31720 -2.192325\n",
"Name: pop_score, dtype: float64\n",
"FN\n",
"track_id\n",
"62450 3.002358\n",
"Name: pop_score, dtype: float64\n",
"===============\n",
"Genre : Rock\n",
"TP\n",
"track_id\n",
"85789 3.288215\n",
"Name: pop_score, dtype: float64\n",
"FP\n",
"track_id\n",
"9520 -2.686833\n",
"Name: pop_score, dtype: float64\n",
"TN\n",
"track_id\n",
"62979 -3.056946\n",
"Name: pop_score, dtype: float64\n",
"FN\n",
"track_id\n",
"114396 2.514736\n",
"Name: pop_score, dtype: float64\n",
"===============\n"
]
}
],
"source": [
"samples = []\n",
"#genres_survey = tracks[\"track\"]['genre_top'].value_counts().index\n",
"genres_survey = ['Electronic', 'Experimental', 'Folk', 'Hip-Hop', 'Instrumental', 'International', 'Pop', 'Rock']\n",
"\n",
"for genre in genres_survey:\n",
" samples.extend(n_samples_genre(1, genre, 0, 0))"
]
},
{
"cell_type": "markdown",
"metadata": {
"id": "U5W-FfDn9cNS"
},
"source": [
"## Samples creation helper"
]
},
{
"cell_type": "code",
"execution_count": 41,
"metadata": {
"id": "w0rucxFu9cNT"
},
"outputs": [],
"source": [
"import shutil\n",
"samples = [str(sample.index.values[0]).zfill(6) for sample in samples]\n",
"\n",
"for sample in samples:\n",
" shutil.copyfile(f\"{datasource}/fma_large/{sample[0:3]}/{sample}.mp3\", f\"{datasource}/survey_samples/{sample}.mp3\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Survey analysis \n",
"\n",
"For all of the 32 samples selected previously, participants are invited to answer to 3 questions after having listened to the corresponding 30s sample:\n",
"1. Would you like to listen to the rest of this music piece ?\n",
"1. Would you recommend this music piece to your friend ?\n",
"1. Would you be surprised to hear this piece in the radio ? \n",
"\n",
"These questions are thought to reflect different dimensions of popularity: the first one is directed toward personal taste and could tell if one would listen casually to the song, the second one add another layer to characterize the piece's capacity to reach broad audiences, and the last one tries to consider one's perception of popularity beside their personal taste. This helps in answering our research question because we defined popularity as how much a song is listened to. With these questions, we will verify - or not - whether this assumptions matches ones's perception of popularity.\n",
"\n",
"\n",
"For each of these questions, answers are collected on a five-level Likert scale proposing the following options:\n",
"- Definitely yes\n",
"- Yes\n",
"- I don't know\n",
"- No\n",
"- Definitely no\n",
"\n",
"Note that in our understanding of popularity, a popular song should receive high scores for the two first question (\"Definitely yes\") but a low score for the last question (\"Definitely no\").\n",
"\n",
"The survey is far from perfect and has some bias : the first two questions are focusing more on the 'why' a music becomes popular, and we only make suppositions. However, if we observe a correlation between actual popularity and positive answers, this would not necessarly mean that one is the cause of the other, and we should keep that in mind. \n",
"\n",
"Only a couple of participants have taken part in this survey so far, hence no results can be drawn. However we already expose the main steps of the analysis pipeline that we intend to use to investigate the results obtained through the survey. \n",
"\n",
"### Loading survey's data"
]
},
{
"cell_type": "code",
"execution_count": 120,
"metadata": {},
"outputs": [],
"source": [
"# path to the folders hosting informations about the survey\n",
"PATH_SURVEY_FOLDER = \"./Musicology/\"\n",
"# general informations\n",
"TASK_FOLDER_ZIP = \"339_music_popularity_task_csv.zip\"\n",
"# participant's answers\n",
"TASK_RUN_FOLDER_ZIP = \"339_music_popularity_task_run_csv.zip\""
]
},
{
"cell_type": "code",
"execution_count": 121,
"metadata": {},
"outputs": [],
"source": [
"# zip with general informations\n",
"zf_task = ZipFile(PATH_SURVEY_FOLDER+TASK_FOLDER_ZIP)\n",
"# zip with answers\n",
"zf_task_run = ZipFile(PATH_SURVEY_FOLDER+TASK_RUN_FOLDER_ZIP)"
]
},
{
"cell_type": "code",
"execution_count": 122,
"metadata": {},
"outputs": [],
"source": [
"# in the code, we'll refer to the above mentionned questions with the following labels\n",
"QUESTIONS = [\"like\", \"recommend\", \"radio\"]"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Create mapping from survey's identifiers to fma's one\n",
"\n",
"Surveys answers are identified by a task id corresponding to one of the selected samples. This id refers to the question on CSZ plateform, which is different from the track identifier. Thus, we recreate the mapping from the task id to the track id thanks to mp3 file associated with the different questions (hence, task id)."
]
},
{
"cell_type": "code",
"execution_count": 123,
"metadata": {},
"outputs": [],
"source": [
"# relink tasks id (samples in the survey) to track ids from fma dataset\n",
"df_task = pd.read_csv(zf_task.open('music_popularity_task.csv'))\n",
"df_task = df_task[[\"id\", \"info_filename\", \"n_answers\"]]\n",
"df_task[\"info_filename\"] = df_task[\"info_filename\"].apply(lambda filemp3: int(filemp3[:-4]))\n",
"df_task.rename({\"info_filename\":\"track_id\"}, axis=1, inplace=True)\n",
"# create a dict -> {task id: track id}\n",
"dict_task_to_track = pd.Series(df_task.track_id.values,index=df_task.id).to_dict()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Load participant's answers\n",
"\n",
"The answers provided by the participant are then loaded and manipulated to make them more convenient to handle.\n",
"Notably, the Likert scale is converted to a linear, equidistant, scale ranging from -2 for \"Definitely no\" to 2 for \"Definitely yes\"."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"df_survey = pd.read_csv(zf_task_run.open('music_popularity_task_run.csv'))\n",
"df_survey = df_survey[[\"id\", \"task_id\", \"user_id\", \"user_ip\", \"info_0\", \"info_1\", \"info_2\"]]\n",
"\n",
"# convert Likert scale to numerical values\n",
"convert_likert_scale = {\"Definitely no\":-2,\n",
" \"No\":-1,\n",
" \"I don't know\":0,\n",
" \"Yes\":1,\n",
" \"Definitely yes\":2\n",
" }\n",
"for col in df_survey.columns:\n",
" if 'info' in col:\n",
" df_survey[col] = df_survey[col].map(convert_likert_scale)\n",
"\n",
"# rename column based on the questions asked\n",
"rename_columns = {\"info_0\":QUESTIONS[0], \"info_1\":QUESTIONS[1], \"info_2\":QUESTIONS[2]}\n",
"df_survey.rename(mapper=rename_columns, inplace=True, axis=1)\n",
"\n",
"# retrieve track id (fma) from task id (CSZ)\n",
"df_survey[\"track_id\"] = df_survey[\"task_id\"].map(dict_task_to_track)\n",
"\n",
"# create unique identifier for the user either based on IP or user id given by the platform and anonymize\n",
"user_identifiers = np.array(list(set(df_survey.user_id.dropna()))+list(set(df_survey.user_ip.dropna())), dtype=str)\n",
"user_anon = np.arange(1,len(user_identifiers)+1, dtype=str)\n",
"dict_anon = dict(zip(user_identifiers, user_anon))\n",
"df_survey[\"user\"] = [str(uid) if str(uid) != \"nan\" \n",
" else str(uip) \n",
" for uid, uip in zip(df_survey.user_id, df_survey.user_ip)\n",
" ]\n",
"df_survey[\"user\"] = df_survey[\"user\"].map(dict_anon)\n",
"df_survey.drop([\"user_id\", \"user_ip\"], axis=1, inplace=True)\n",
"\n",
"df_survey.sample(5)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Compute multi-level modeling variables\n",
"\n",
"As survey bear intrinsic bias and answers are highly influenced by the respondent personality, we compute multi-level variable that rescale one participant answers around their mean value for each of the 3 questions. "
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"users_means = df_survey.groupby(\"user\").mean()\n",
"users_means"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"for q in QUESTIONS:\n",
" df_survey[\"rel_\"+q] = [score-users_means.loc[user_id][q]\n",
" for (score, user_id) in zip(df_survey[q], df_survey[\"user\"])\n",
" ]\n",
" \n",
"df_survey.sample(5)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Enrich with fma and previously computed features\n",
"\n",
"The DF containing the surveys answers are then enriched with fma's informations as well as previously computed popularity features so that all following analysis can be solely based on this DataFrame. "
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"samples_id = [sample.index[0] for sample in samples]\n",
"additional_info = pop_df.loc[samples_id].drop(\"date_created\", axis=1)\n",
"\n",
"# retrieve type of classification (TP,FP,TN,FN) from the way samples were generated\n",
"additional_info[\"type\"] = [\"tp\" if (i%4)==0\n",
" else \"fp\" if (i%4)==1\n",
" else \"tn\" if (i%4)==2\n",
" else \"fn\"\n",
" for i in np.arange(len(additional_info))\n",
" ]\n",
"\n",
"# add these informations in the survey DF\n",
"df_survey = df_survey.merge(additional_info, on=\"track_id\")\n",
"df_survey.sample(5)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Results analysis\n",
"\n",
"#### Inter-rater agreement\n",
"\n",
"\n",
"First thing, we can assess the inter-rater agreement thanks to several methods developped along the second half of the XXth century that are pre-implemented in `nltk`'s metrics package. These first measures can bring informations regarding the shared (or not) behaviors or perception of the participants pool (eg. a value high agreement measure, close to 1, regarding the radio question could indicate that the perception of what kind of music pieces are played on the radio is shared amongst participants).\n",
"\n",
"Metrics computed below are Cohen's $\\kappa$ [(Cohen, 1960)](https://journals.sagepub.com/doi/10.1177/001316446002000104), Fleiss' $\\kappa$ [(Fleiss, 1971)](http://www.wpic.pitt.edu/research/biometrics/Publications/Biometrics%20Archives%20PDF/395-1971%20Fleiss0001.pdf) and Krippendorff's $\\alpha$ [(Krippendorff, 1989)](https://repository.upenn.edu/cgi/viewcontent.cgi?article=1232&context=asc_papers).\n",
"\n",
"- Cohen's $\\kappa$ measures agreement between raters while taking into account the probability that they could agree \"by chance\". It is computed as $$ \\kappa = \\frac{p_o-p_e}{1-p_e} $$ where $p_o$ is the relative observed agreement between raters and $p_e$ is the probability of agreement by chance. For $N>2$ raters, $\\kappa_{ij}$ is computed for each pair $(i,j)$ of raters and the output value is a naive average over the set of $\\{\\kappa_{ij}\\}_{i,j=1}^{N}$.\n",
"- Fleiss' $\\kappa$ is computed similarly to Cohen's $\\kappa$ but taking into account averages for both the expected and observed agreement for each pair of raters.\n",
"\n",
"Both of these measures range from -1 to 1 and it is commonly accepted (see [MiniTab online documentation](https://support.minitab.com/en-us/minitab/18/help-and-how-to/quality-and-process-improvement/measurement-system-analysis/how-to/attribute-agreement-analysis/attribute-agreement-analysis/interpret-the-results/all-statistics-and-graphs/kappa-statistics/)) that $\\alpha<0$ indicates more disagreement than what could be expected by chance, $\\alpha=0$ agreement is the same that what could be expected by chance, and substantial agreement can be underlined above $\\alpha=0.6$.\n",
"\n",
"- Krippendorff's $\\alpha$ was developped to be more suitable to diverse types of ratings and data aquired. It is expressed as $$ \\alpha = 1- \\frac{D_o}{D_e} $$ where $D_o$ is the observed disagreement and $D_e$ is the one expected by chance. Those are computed as the weighted average dispersion of the ratings."
]
},
{
"cell_type": "code",
"execution_count": 189,
"metadata": {},
"outputs": [],
"source": [
"participants_list = list(set(df_survey.user))\n",
"\n",
"# should be unique so sum is just the score given by the rater\n",
"survey_grouped_user_track = df_survey.groupby([\"user\", \"track_id\"]).sum()"
]
},
{
"cell_type": "code",
"execution_count": 210,
"metadata": {},
"outputs": [],
"source": [
"def dist_labels(l1, l2):\n",
" \"\"\" define distance between labels (equidistant and symmetric) \"\"\"\n",
" return np.abs(l1-l2)"
]
},
{
"cell_type": "code",
"execution_count": 208,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"===============\n",
"Feature : like\n",
"Agreement measures:\n",
"kappa : -0.277\n",
"fleiss : -0.277\n",
"alpha : 0.30859\n",
"scotts : -0.34988\n",
"===============\n",
"Feature : recommend\n",
"Agreement measures:\n",
"kappa : -0.04474\n",
"fleiss : -0.04474\n",
"alpha : 0.41244\n",
"scotts : -0.076636\n",
"===============\n",
"Feature : radio\n",
"Agreement measures:\n",
"kappa : -0.44934\n",
"fleiss : -0.44934\n",
"alpha : 0.18294\n",
"scotts : -0.53342\n",
"===============\n",
"Feature : rel_like\n",
"Agreement measures:\n",
"kappa : 0.078125\n",
"fleiss : 0.078125\n",
"alpha : 0.39393\n",
"scotts : -0.058296\n",
"===============\n",
"Feature : rel_recommend\n",
"Agreement measures:\n",
"kappa : 0.080078\n",
"fleiss : 0.080078\n",
"alpha : 0.36065\n",
"scotts : -0.045505\n",
"===============\n",
"Feature : rel_radio\n",
"Agreement measures:\n",
"kappa : -0.1875\n",
"fleiss : -0.1875\n",
"alpha : 0.19036\n",
"scotts : -0.37324\n"
]
}
],
"source": [
"for q in QUESTIONS+[\"rel_\"+q for q in QUESTIONS]:\n",
" print(\"===============\")\n",
" print(\"Feature : \", q)\n",
"\n",
" taskdata = []\n",
" for i, participant in enumerate(participants_list):\n",
" # convert data to nltk required format\n",
" taskdata = taskdata+[[i\n",
" ,str(samples_id[j])\n",
" ,survey_grouped_user_track.loc[participant][q].loc[samples_id[j]]\n",
" ]\n",
" for j in range(0,len(samples_id))\n",
" ]\n",
"\n",
" ratingtask = agreement.AnnotationTask(data=taskdata, distance=dist_labels)\n",
" print(\"Agreement measures:\")\n",
" print(\"kappa : {0:.5g}\".format(ratingtask.kappa()))\n",
" print(\"fleiss : {0:.5g}\".format(ratingtask.multi_kappa()))\n",
" print(\"alpha : {0:.5g}\".format(ratingtask.alpha()))\n",
" #print(\"scotts : {0:.5g}\".format(ratingtask.pi()))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Correlation analysis\n",
"\n",
"Then, correlations are studied. We will look for possible significant (based on $p$ values) correlations between the average scores given by the participant for the different questions but also between these human features and the features used to define popularity (number of streams, likes, comments for song and album) and the popularity score itself."
]
},
{
"cell_type": "code",
"execution_count": 211,
"metadata": {},
"outputs": [],
"source": [
"def corr_sig(df):\n",
" \"\"\" return matrix of p values of df based on its correlation matrix \"\"\"\n",
" p_matrix = np.zeros(shape=(df.shape[1],df.shape[1]))\n",
" for col in df.columns:\n",
" for col2 in df.drop(col,axis=1).columns:\n",
" _ , p = stats.pearsonr(df[col],df[col2])\n",
" p_matrix[df.columns.to_list().index(col),df.columns.to_list().index(col2)] = p\n",
" return p_matrix"
]
},
{
"cell_type": "code",
"execution_count": 216,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"
"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"df_survey_pop = df_survey[QUESTIONS\n",
" +[\"rel_\"+q for q in QUESTIONS]\n",
" +POP_FEATURES\n",
" +[\"album_\"+f for f in POP_FEATURES]\n",
" +[\"pop_score\"]\n",
" ]\n",
"\n",
"corr_survey = df_survey_pop.corr()\n",
"\n",
"p_values = corr_sig(df_survey_pop) # compute p-values\n",
"mask_sig = np.invert(np.tril(p_values<0.05)) # mask non significant correlations\n",
"\n",
"fig, ax = plt.subplots(figsize=(15,12))\n",
"ax = sns.heatmap(corr_survey\n",
" #, mask=mask_sig # mask non significant values\n",
" , square=True\n",
" , annot=True\n",
" , fmt=\".2f\"\n",
" , vmin=-1\n",
" , vmax=1\n",
" , cmap=\"RdBu\"\n",
" )\n",
"plt.show()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"The differences between the four selected groups (true positive, false postivie, flase negative and true negative) can also be explored. This could highlight interesting insights on how human perceive the samples classified as popular and possibly on what kind of samples are misclassified.\n",
"For example are True Positive cases given high liking and recommending scores and low surprises on their potentiality to be played on the radio by raters ? Also, are False Negative (Positive) cases rated by participants more similarly as True Positive or True Negative ?"
]
},
{
"cell_type": "code",
"execution_count": 217,
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"