{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# This process is for the country level, not for the frontline labels" ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "/Users/barr001/anaconda3/envs/python3.7/lib/python3.7/site-packages/statsmodels/tools/_testing.py:19: FutureWarning: pandas.util.testing is deprecated. Use the functions in the public API at pandas.testing instead.\n", " import pandas.util.testing as tm\n" ] } ], "source": [ "import pandas as pd\n", "import numpy as np\n", "from datetime import datetime, timedelta \n", "import os\n", "import seaborn as sns\n", "import matplotlib.pyplot as plt\n", "import glob" ] }, { "cell_type": "code", "execution_count": 35, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "(3471, 8)" ] }, "execution_count": 35, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# Bring in the data\n", "# This is the data I pulled from the finance_country_level.sql queries\n", "old = pd.read_csv('/Users/barr001/OneDrive - Sony Music Entertainment/Desktop/Finance Market Share Trending/mx_oldest_data_thru_aug_21.csv')\n", "df = pd.read_csv('/Users/barr001/OneDrive - Sony Music Entertainment/Desktop/Finance Market Share Trending/mx_data_thru_aug_21.csv')\n", "df.shape" ] }, { "cell_type": "code", "execution_count": 36, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
release_monthmonth_end_datemonth_short_nameisrc_countstreamsmonths_after_releasecountry_codesix_month_segmentsegment_min_monthmonths_after_min
02014-08-312015-03-31MAR 2015177960636907MX142014-05-0110
12014-06-302017-09-30SEP 20175807603074739MX142014-05-0140
22014-05-312019-12-31DEC 201929385482946917267MX142014-05-0167
32014-09-302020-07-31JUL 20202523873717970MX142014-05-0174
42014-05-312014-07-31JUL 20142467221455519662MX142014-05-012
.................................
4302014-05-312014-10-31OCT 20142546731793542235MX142014-05-015
4312014-05-312019-04-30APR 201929236268048884259MX142014-05-0159
4322014-09-302021-02-28FEB 20212292849771777MX142014-05-0181
4332014-08-312016-09-30SEP 20161811809922725MX142014-05-0128
4342014-09-302020-12-31DEC 20202324816828675MX142014-05-0179
\n", "

435 rows × 10 columns

\n", "
" ], "text/plain": [ " release_month month_end_date month_short_name isrc_count streams \\\n", "0 2014-08-31 2015-03-31 MAR 2015 1779 6063690 \n", "1 2014-06-30 2017-09-30 SEP 2017 5807 6030747 \n", "2 2014-05-31 2019-12-31 DEC 2019 293854 829469172 \n", "3 2014-09-30 2020-07-31 JUL 2020 2523 8737179 \n", "4 2014-05-31 2014-07-31 JUL 2014 246722 145551966 \n", ".. ... ... ... ... ... \n", "430 2014-05-31 2014-10-31 OCT 2014 254673 179354223 \n", "431 2014-05-31 2019-04-30 APR 2019 292362 680488842 \n", "432 2014-09-30 2021-02-28 FEB 2021 2292 8497717 \n", "433 2014-08-31 2016-09-30 SEP 2016 1811 8099227 \n", "434 2014-09-30 2020-12-31 DEC 2020 2324 8168286 \n", "\n", " months_after_release country_code six_month_segment segment_min_month \\\n", "0 7 MX 14 2014-05-01 \n", "1 39 MX 14 2014-05-01 \n", "2 67 MX 14 2014-05-01 \n", "3 70 MX 14 2014-05-01 \n", "4 2 MX 14 2014-05-01 \n", ".. ... ... ... ... \n", "430 5 MX 14 2014-05-01 \n", "431 59 MX 14 2014-05-01 \n", "432 77 MX 14 2014-05-01 \n", "433 25 MX 14 2014-05-01 \n", "434 75 MX 14 2014-05-01 \n", "\n", " months_after_min \n", "0 10 \n", "1 40 \n", "2 67 \n", "3 74 \n", "4 2 \n", ".. ... \n", "430 5 \n", "431 59 \n", "432 81 \n", "433 28 \n", "434 79 \n", "\n", "[435 rows x 10 columns]" ] }, "execution_count": 36, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# Adding some columns to match up with the newer data\n", "old['six_month_segment']=14\n", "del old['partner_name']\n", "old['release_month'] = pd.to_datetime(old['release_month'])\n", "old['month_end_date'] = pd.to_datetime(old['month_end_date'])\n", "old['segment_min_month'] = '2014-05-01'\n", "old['segment_min_month'] = pd.to_datetime(old['segment_min_month'])\n", "old['months_after_min']=(old['month_end_date'].dt.to_period('M')).astype('int32')-old['segment_min_month'].dt.to_period('M').astype('int32')\n", "old" ] }, { "cell_type": "code", "execution_count": 37, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
release_monthmonth_end_datemonth_short_nameisrc_countstreamsmonths_after_releasecountry_code
02016-09-302017-11-30NOV 201731971654217914MX
12016-06-302018-01-31JAN 201824371280477419MX
22015-06-302018-03-31MAR 201842561282017733MX
32015-09-302018-12-31DEC 20183085422662239MX
42017-03-312019-01-31JAN 201934731316245722MX
........................
34662015-11-302019-07-31JUL 20194560819216344MX
34672014-12-312019-10-31OCT 20191837857315258MX
34682019-10-312019-12-31DEC 20194094615332722MX
34692017-12-312020-12-31DEC 20202031334996436MX
34702016-11-302021-06-30JUN 20212310248344855MX
\n", "

3471 rows × 7 columns

\n", "
" ], "text/plain": [ " release_month month_end_date month_short_name isrc_count streams \\\n", "0 2016-09-30 2017-11-30 NOV 2017 3197 16542179 \n", "1 2016-06-30 2018-01-31 JAN 2018 2437 12804774 \n", "2 2015-06-30 2018-03-31 MAR 2018 4256 12820177 \n", "3 2015-09-30 2018-12-31 DEC 2018 3085 4226622 \n", "4 2017-03-31 2019-01-31 JAN 2019 3473 13162457 \n", "... ... ... ... ... ... \n", "3466 2015-11-30 2019-07-31 JUL 2019 4560 8192163 \n", "3467 2014-12-31 2019-10-31 OCT 2019 1837 8573152 \n", "3468 2019-10-31 2019-12-31 DEC 2019 4094 61533272 \n", "3469 2017-12-31 2020-12-31 DEC 2020 2031 3349964 \n", "3470 2016-11-30 2021-06-30 JUN 2021 2310 2483448 \n", "\n", " months_after_release country_code \n", "0 14 MX \n", "1 19 MX \n", "2 33 MX \n", "3 39 MX \n", "4 22 MX \n", "... ... ... \n", "3466 44 MX \n", "3467 58 MX \n", "3468 2 MX \n", "3469 36 MX \n", "3470 55 MX \n", "\n", "[3471 rows x 7 columns]" ] }, "execution_count": 37, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# Could pull without the partner name field, just left that in case you generalize to more partners in the future\n", "del df['partner_name']\n", "df" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### For this model, I initially included all data more than 42 months old, but this would bring in more data every month (as new tracks would become 42 months old every month), but that would have some overlap with the under 42 month model\n", "##### That's why I will calculate the segment_min_month below, so every month in the 6 month segment will be treated as one" ] }, { "cell_type": "code", "execution_count": 38, "metadata": {}, "outputs": [], "source": [ "# Adding the 6 month segment to the data, which I need to add segment min_date\n", "# I'll use segment min_date to make sure I'm including static data that won't change every month\n", "# I know this is a bit manual, feel free to update if you have a better way of doing this!\n", "df['release_month'] = pd.to_datetime(df['release_month'])\n", "df['month_end_date'] = pd.to_datetime(df['month_end_date'])\n", "df['six_month_segment']=0\n", "for x in range(len(df)):\n", " if df.iloc[x,0]>=pd.Timestamp('2021-04-01'):\n", " df.iloc[x,7]=0\n", " elif df.iloc[x,0]>=pd.Timestamp('2020-10-01'):\n", " df.iloc[x,7]=1\n", " elif df.iloc[x,0]>=pd.Timestamp('2020-04-01'):\n", " df.iloc[x,7]=2\n", " elif df.iloc[x,0]>=pd.Timestamp('2019-10-01'):\n", " df.iloc[x,7]=3\n", " elif df.iloc[x,0]>=pd.Timestamp('2019-04-01'):\n", " df.iloc[x,7]=4\n", " elif df.iloc[x,0]>=pd.Timestamp('2018-10-01'):\n", " df.iloc[x,7]=5\n", " elif df.iloc[x,0]>=pd.Timestamp('2018-04-01'):\n", " df.iloc[x,7]=6\n", " elif df.iloc[x,0]>=pd.Timestamp('2017-10-01'):\n", " df.iloc[x,7]=7\n", " elif df.iloc[x,0]>=pd.Timestamp('2017-04-01'):\n", " df.iloc[x,7]=8\n", " elif df.iloc[x,0]>=pd.Timestamp('2016-10-01'):\n", " df.iloc[x,7]=9\n", " elif df.iloc[x,0]>=pd.Timestamp('2016-04-01'):\n", " df.iloc[x,7]=10\n", " elif df.iloc[x,0]>=pd.Timestamp('2015-10-01'):\n", " df.iloc[x,7]=11\n", " elif df.iloc[x,0]>=pd.Timestamp('2015-04-01'):\n", " df.iloc[x,7]=12\n", " elif df.iloc[x,0]>=pd.Timestamp('2014-10-01'):\n", " df.iloc[x,7]=13\n", " else:\n", " df.iloc[x,7]=100" ] }, { "cell_type": "code", "execution_count": 39, "metadata": {}, "outputs": [], "source": [ "# Now that I have the six_month_segment, I can caluclate the min month for each segment \n", "# The min month is the earliest month in each release segment\n", "# So if a segment goes from October 2020 to March 2021, the segment_min_month is October 2020\n", "df['segment_min_month']=''\n", "for x in range(len(df)):\n", " if df.iloc[x,7]==0:\n", " df.iloc[x,8]='2021-04-01' \n", " elif df.iloc[x,7]==1:\n", " df.iloc[x,8]='2020-10-01' \n", " elif df.iloc[x,7]==2:\n", " df.iloc[x,8]='2020-04-01'\n", " elif df.iloc[x,7]==3:\n", " df.iloc[x,8]='2019-10-01'\n", " elif df.iloc[x,7]==4:\n", " df.iloc[x,8]='2019-04-01'\n", " elif df.iloc[x,7]==5:\n", " df.iloc[x,8]='2018-10-01'\n", " elif df.iloc[x,7]==6:\n", " df.iloc[x,8]='2018-04-01'\n", " elif df.iloc[x,7]==7:\n", " df.iloc[x,8]='2017-10-01'\n", " elif df.iloc[x,7]==8:\n", " df.iloc[x,8]='2017-04-01'\n", " elif df.iloc[x,7]==9:\n", " df.iloc[x,8]='2016-10-01'\n", " elif df.iloc[x,7]==10:\n", " df.iloc[x,8]='2016-04-01'\n", " elif df.iloc[x,7]==11:\n", " df.iloc[x,8]='2015-10-01'\n", " elif df.iloc[x,7]==12:\n", " df.iloc[x,8]='2015-04-01'\n", " elif df.iloc[x,7]==13:\n", " df.iloc[x,8]='2014-10-01'\n", " else:\n", " df.iloc[x,8]='2014-01-01'" ] }, { "cell_type": "code", "execution_count": 40, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "release_month datetime64[ns]\n", "month_end_date datetime64[ns]\n", "month_short_name object\n", "isrc_count int64\n", "streams int64\n", "months_after_release int64\n", "country_code object\n", "six_month_segment int64\n", "segment_min_month datetime64[ns]\n", "dtype: object" ] }, "execution_count": 40, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df['segment_min_month'] = pd.to_datetime(df['segment_min_month'])\n", "df.dtypes" ] }, { "cell_type": "code", "execution_count": 41, "metadata": {}, "outputs": [], "source": [ "# Now I calculate the months_after_min field to use to simply filter the data\n", "# This is generalized, so you can alwayd use the same number instead of having to update segment_min_month dates every year\n", "df['months_after_min']=(df['month_end_date'].dt.to_period('M')).astype('int32')-df['segment_min_month'].dt.to_period('M').astype('int32')" ] }, { "cell_type": "code", "execution_count": 42, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "release_month datetime64[ns]\n", "month_end_date datetime64[ns]\n", "month_short_name object\n", "isrc_count int64\n", "streams int64\n", "months_after_release int64\n", "country_code object\n", "six_month_segment int64\n", "segment_min_month datetime64[ns]\n", "months_after_min int64\n", "dtype: object" ] }, "execution_count": 42, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# Combining the older tracks with the newer ones\n", "combined_no_label = pd.concat([df,old])\n", "combined_no_label['release_month'] = pd.to_datetime(combined_no_label['release_month'])\n", "combined_no_label['month_end_date'] = pd.to_datetime(combined_no_label['month_end_date'])\n", "combined_no_label.dtypes\n", "# combined_no_label.shape" ] }, { "cell_type": "code", "execution_count": 43, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
release_monthmonth_end_datemonth_short_nameisrc_countstreamsmonths_after_releasecountry_codesix_month_segmentsegment_min_monthmonths_after_min
02016-09-302017-11-30NOV 201731971654217914MX102016-04-0119
12016-06-302018-01-31JAN 201824371280477419MX102016-04-0121
22015-06-302018-03-31MAR 201842561282017733MX122015-04-0135
32015-09-302018-12-31DEC 20183085422662239MX122015-04-0144
42017-03-312019-01-31JAN 201934731316245722MX92016-10-0127
\n", "
" ], "text/plain": [ " release_month month_end_date month_short_name isrc_count streams \\\n", "0 2016-09-30 2017-11-30 NOV 2017 3197 16542179 \n", "1 2016-06-30 2018-01-31 JAN 2018 2437 12804774 \n", "2 2015-06-30 2018-03-31 MAR 2018 4256 12820177 \n", "3 2015-09-30 2018-12-31 DEC 2018 3085 4226622 \n", "4 2017-03-31 2019-01-31 JAN 2019 3473 13162457 \n", "\n", " months_after_release country_code six_month_segment segment_min_month \\\n", "0 14 MX 10 2016-04-01 \n", "1 19 MX 10 2016-04-01 \n", "2 33 MX 12 2015-04-01 \n", "3 39 MX 12 2015-04-01 \n", "4 22 MX 9 2016-10-01 \n", "\n", " months_after_min \n", "0 19 \n", "1 21 \n", "2 35 \n", "3 44 \n", "4 27 " ] }, "execution_count": 43, "metadata": {}, "output_type": "execute_result" } ], "source": [ "combined_no_label.head()" ] }, { "cell_type": "code", "execution_count": 44, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
report_monthIndustry StreamsSME Streams
822020-11-3090982409211801623284
832020-12-3196984570841857614761
842021-01-3195702046711816356994
852021-02-2891698945631748483924
862021-03-31104864073221992613245
\n", "
" ], "text/plain": [ " report_month Industry Streams SME Streams\n", "82 2020-11-30 9098240921 1801623284\n", "83 2020-12-31 9698457084 1857614761\n", "84 2021-01-31 9570204671 1816356994\n", "85 2021-02-28 9169894563 1748483924\n", "86 2021-03-31 10486407322 1992613245" ] }, "execution_count": 44, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# Replace this with the correct country/dates market share\n", "# be sure to filter to only include the months you want to predict on (in this case up to March 2021)\n", "ms = pd.read_excel(\"/Users/barr001/Downloads/Spotify Streams thru Aug'21 US GB BR MX.xlsx\", skiprows=2, sheet_name='MX')\n", "ms = ms.rename(columns={'Unnamed: 0': 'report_month'})\n", "ms = ms[['report_month','Industry Streams', 'SME Streams']]\n", "ms['report_month'] = ms['report_month']-pd.offsets.MonthEnd(0)\n", "ms['report_month'] = pd.to_datetime(ms['report_month'])\n", "ms = ms[ms['report_month']<='2021-03-31']\n", "ms.tail()" ] }, { "cell_type": "code", "execution_count": 45, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
release_monthmonth_end_datemonth_short_nameisrc_countstreamsmonths_after_releasecountry_codesix_month_segmentsegment_min_monthmonths_after_minreport_monthIndustry StreamsSME Streams
02016-09-302017-11-30NOV 201731971654217914MX102016-04-01192017-11-303844132455981814600
12015-01-312017-11-30NOV 20171641475653734MX132014-10-01372017-11-303844132455981814600
22017-10-312017-11-30NOV 20173399245422721MX72017-10-0112017-11-303844132455981814600
32015-04-302017-11-30NOV 20172655463842831MX122015-04-01312017-11-303844132455981814600
42015-12-312017-11-30NOV 20172367754627323MX112015-10-01252017-11-303844132455981814600
..........................................
34812014-07-312014-08-31AUG 2014156520609241MX142014-05-0132014-08-31597401155158614552
34822014-05-312014-08-31AUG 20142512081585829373MX142014-05-0132014-08-31597401155158614552
34832014-06-302014-06-30JUN 2014399117107210MX142014-05-0112014-06-30453775292122691583
34842014-05-312014-06-30JUN 20142399371293180771MX142014-05-0112014-06-30453775292122691583
34852014-05-312014-05-31MAY 20143745888319354510MX142014-05-0102014-05-31452032051122136482
\n", "

3486 rows × 13 columns

\n", "
" ], "text/plain": [ " release_month month_end_date month_short_name isrc_count streams \\\n", "0 2016-09-30 2017-11-30 NOV 2017 3197 16542179 \n", "1 2015-01-31 2017-11-30 NOV 2017 1641 4756537 \n", "2 2017-10-31 2017-11-30 NOV 2017 3399 24542272 \n", "3 2015-04-30 2017-11-30 NOV 2017 2655 4638428 \n", "4 2015-12-31 2017-11-30 NOV 2017 2367 7546273 \n", "... ... ... ... ... ... \n", "3481 2014-07-31 2014-08-31 AUG 2014 1565 2060924 \n", "3482 2014-05-31 2014-08-31 AUG 2014 251208 158582937 \n", "3483 2014-06-30 2014-06-30 JUN 2014 3991 1710721 \n", "3484 2014-05-31 2014-06-30 JUN 2014 239937 129318077 \n", "3485 2014-05-31 2014-05-31 MAY 2014 374588 831935451 \n", "\n", " months_after_release country_code six_month_segment segment_min_month \\\n", "0 14 MX 10 2016-04-01 \n", "1 34 MX 13 2014-10-01 \n", "2 1 MX 7 2017-10-01 \n", "3 31 MX 12 2015-04-01 \n", "4 23 MX 11 2015-10-01 \n", "... ... ... ... ... \n", "3481 1 MX 14 2014-05-01 \n", "3482 3 MX 14 2014-05-01 \n", "3483 0 MX 14 2014-05-01 \n", "3484 1 MX 14 2014-05-01 \n", "3485 0 MX 14 2014-05-01 \n", "\n", " months_after_min report_month Industry Streams SME Streams \n", "0 19 2017-11-30 3844132455 981814600 \n", "1 37 2017-11-30 3844132455 981814600 \n", "2 1 2017-11-30 3844132455 981814600 \n", "3 31 2017-11-30 3844132455 981814600 \n", "4 25 2017-11-30 3844132455 981814600 \n", "... ... ... ... ... \n", "3481 3 2014-08-31 597401155 158614552 \n", "3482 3 2014-08-31 597401155 158614552 \n", "3483 1 2014-06-30 453775292 122691583 \n", "3484 1 2014-06-30 453775292 122691583 \n", "3485 0 2014-05-31 452032051 122136482 \n", "\n", "[3486 rows x 13 columns]" ] }, "execution_count": 45, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# Merge the stream data with the market share data\n", "old_grouped = combined_no_label.merge(ms, left_on='month_end_date', right_on='report_month')\n", "old_grouped" ] }, { "cell_type": "code", "execution_count": 46, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
month_end_dateIndustry StreamsSME Streamsisrc_countstreamsmonths_after_releasesix_month_segmentmonths_after_minmarket_share
02018-05-3147428014391153036250304373585805848230702400.123515
12018-06-3046765757681134772662297532576292729235702450.123230
22018-07-3149106555801165177617300284606977501240702500.123604
32018-08-3152015603151243729638305942656867039245702550.126283
42018-09-3051812594941225264103303000647684076250702600.125005
52018-10-31550923921712832235473185617208541205281485530.130845
62018-11-30547177036412395009723153736922599195391485640.126515
72018-12-31574877193812918294033157057127547925501485750.123984
82019-01-31597594038513277357263186087312767685611485860.122370
92019-02-28584986198312812579093160547160920325721485970.122412
102019-03-31660698861214485017453213898089405565831486080.122437
112019-04-30626179438013728801763360027922025168672209070.126514
122019-05-31670124638714724098043382768475639708842209240.126479
132019-06-30674728625214729260393393708360063709012209410.123903
142019-07-31685092807314880002403373038610032149182209580.125677
152019-08-31718742851615720555133370158997745029352209750.125187
162019-09-30722239295215863275523378449086860779522209920.125815
172019-10-31766598343817036208003652781044299639124228612970.136225
182019-11-3075700722481629203870364958974778279126528613200.128767
192019-12-3179938523221675126969362780993757445128828613430.124315
202020-01-31828835287417075529153644731005543010131128613660.121320
212020-02-29833896892316939497923593601008302178133428613890.120914
222020-03-31873844104816803632403607771011795136135728614120.115787
232020-04-3078865897181536801817375102959080779165334617230.121609
242020-05-31841371973516380633213809141024619843168234617520.121780
252020-06-30830114587216230396293751631016292293171134617810.122428
262020-07-31871349743217183065263764231077726787174034618100.123685
272020-08-31893943186317863554143731071086640938176934618390.121556
282020-09-30881188007117702085433714291060131712179834618680.120307
292020-10-31944142833818926210423857711191634347210040021850.126213
302020-11-30909824092118016232843846611123423410213540022200.123477
312020-12-31969845708418576147613864431142828011217040022550.117836
322021-01-31957020467118163569943856351117318820220540022900.116750
332021-02-28916989456317484839243785811086925023224040023250.118532
342021-03-311048640732219926132453825981230750811227540023600.117366
\n", "
" ], "text/plain": [ " month_end_date Industry Streams SME Streams isrc_count streams \\\n", "0 2018-05-31 4742801439 1153036250 304373 585805848 \n", "1 2018-06-30 4676575768 1134772662 297532 576292729 \n", "2 2018-07-31 4910655580 1165177617 300284 606977501 \n", "3 2018-08-31 5201560315 1243729638 305942 656867039 \n", "4 2018-09-30 5181259494 1225264103 303000 647684076 \n", "5 2018-10-31 5509239217 1283223547 318561 720854120 \n", "6 2018-11-30 5471770364 1239500972 315373 692259919 \n", "7 2018-12-31 5748771938 1291829403 315705 712754792 \n", "8 2019-01-31 5975940385 1327735726 318608 731276768 \n", "9 2019-02-28 5849861983 1281257909 316054 716092032 \n", "10 2019-03-31 6606988612 1448501745 321389 808940556 \n", "11 2019-04-30 6261794380 1372880176 336002 792202516 \n", "12 2019-05-31 6701246387 1472409804 338276 847563970 \n", "13 2019-06-30 6747286252 1472926039 339370 836006370 \n", "14 2019-07-31 6850928073 1488000240 337303 861003214 \n", "15 2019-08-31 7187428516 1572055513 337015 899774502 \n", "16 2019-09-30 7222392952 1586327552 337844 908686077 \n", "17 2019-10-31 7665983438 1703620800 365278 1044299639 \n", "18 2019-11-30 7570072248 1629203870 364958 974778279 \n", "19 2019-12-31 7993852322 1675126969 362780 993757445 \n", "20 2020-01-31 8288352874 1707552915 364473 1005543010 \n", "21 2020-02-29 8338968923 1693949792 359360 1008302178 \n", "22 2020-03-31 8738441048 1680363240 360777 1011795136 \n", "23 2020-04-30 7886589718 1536801817 375102 959080779 \n", "24 2020-05-31 8413719735 1638063321 380914 1024619843 \n", "25 2020-06-30 8301145872 1623039629 375163 1016292293 \n", "26 2020-07-31 8713497432 1718306526 376423 1077726787 \n", "27 2020-08-31 8939431863 1786355414 373107 1086640938 \n", "28 2020-09-30 8811880071 1770208543 371429 1060131712 \n", "29 2020-10-31 9441428338 1892621042 385771 1191634347 \n", "30 2020-11-30 9098240921 1801623284 384661 1123423410 \n", "31 2020-12-31 9698457084 1857614761 386443 1142828011 \n", "32 2021-01-31 9570204671 1816356994 385635 1117318820 \n", "33 2021-02-28 9169894563 1748483924 378581 1086925023 \n", "34 2021-03-31 10486407322 1992613245 382598 1230750811 \n", "\n", " months_after_release six_month_segment months_after_min market_share \n", "0 230 70 240 0.123515 \n", "1 235 70 245 0.123230 \n", "2 240 70 250 0.123604 \n", "3 245 70 255 0.126283 \n", "4 250 70 260 0.125005 \n", "5 528 148 553 0.130845 \n", "6 539 148 564 0.126515 \n", "7 550 148 575 0.123984 \n", "8 561 148 586 0.122370 \n", "9 572 148 597 0.122412 \n", "10 583 148 608 0.122437 \n", "11 867 220 907 0.126514 \n", "12 884 220 924 0.126479 \n", "13 901 220 941 0.123903 \n", "14 918 220 958 0.125677 \n", "15 935 220 975 0.125187 \n", "16 952 220 992 0.125815 \n", "17 1242 286 1297 0.136225 \n", "18 1265 286 1320 0.128767 \n", "19 1288 286 1343 0.124315 \n", "20 1311 286 1366 0.121320 \n", "21 1334 286 1389 0.120914 \n", "22 1357 286 1412 0.115787 \n", "23 1653 346 1723 0.121609 \n", "24 1682 346 1752 0.121780 \n", "25 1711 346 1781 0.122428 \n", "26 1740 346 1810 0.123685 \n", "27 1769 346 1839 0.121556 \n", "28 1798 346 1868 0.120307 \n", "29 2100 400 2185 0.126213 \n", "30 2135 400 2220 0.123477 \n", "31 2170 400 2255 0.117836 \n", "32 2205 400 2290 0.116750 \n", "33 2240 400 2325 0.118532 \n", "34 2275 400 2360 0.117366 " ] }, "execution_count": 46, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# IMPORTANT STEP - filter for only tracks with months_after_min>=48\n", "# This means that we are only including tracks over 42 months old but it will not update every month\n", "# months_after_min>=48 means that all tracks will be at least 43 months old \n", "# (the most recent segment would have tracks ages 48, 47, 46, 45, 44, and 43 months old)\n", "combo_group = old_grouped[old_grouped['months_after_min']>=48].groupby(['month_end_date','Industry Streams','SME Streams']).sum().reset_index()\n", "combo_group['market_share'] = combo_group['streams']/combo_group['Industry Streams']\n", "combo_group" ] }, { "cell_type": "code", "execution_count": 47, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "INFO:fbprophet:Disabling weekly seasonality. Run prophet with weekly_seasonality=True to override this.\n", "INFO:fbprophet:Disabling daily seasonality. Run prophet with daily_seasonality=True to override this.\n" ] }, { "data": { "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# This is the actual model. I'm not doing much optimization, feel free to play around\n", "from fbprophet import Prophet\n", "older_42 = combo_group[['month_end_date', 'market_share']]\n", "older_42 = older_42.rename(columns={'month_end_date': 'ds', 'market_share': 'y'})\n", "monthly_model = Prophet(interval_width=0.95,seasonality_mode='multiplicative')\n", "monthly_model.fit(older_42)\n", "\n", "monthly_forecast = monthly_model.make_future_dataframe(periods=12*1, freq='M')\n", "monthly_forecast = monthly_model.predict(monthly_forecast)\n", "\n", "plt.figure(figsize=(18, 6))\n", "monthly_model.plot(monthly_forecast, xlabel = 'Month', ylabel = 'Market Share')\n", "plt.title('Forecast - Tracks Over 42 Months');" ] }, { "cell_type": "code", "execution_count": 48, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
dsyhat
02018-05-310.124069
12018-06-300.123136
22018-07-310.124375
32018-08-310.124707
42018-09-300.124281
52018-10-310.132125
62018-11-300.127252
72018-12-310.123648
82019-01-310.121993
92019-02-280.121721
102019-03-310.122539
112019-04-300.126377
122019-05-310.126275
132019-06-300.124292
142019-07-310.125504
152019-08-310.126429
162019-09-300.126236
172019-10-310.135017
182019-11-300.127644
192019-12-310.123715
202020-01-310.120574
212020-02-290.122055
222020-03-310.117438
232020-04-300.121737
242020-05-310.121417
252020-06-300.122117
262020-07-310.123075
272020-08-310.121882
282020-09-300.120617
292020-10-310.126160
302020-11-300.123886
312020-12-310.118808
322021-01-310.117914
332021-02-280.118066
342021-03-310.115560
352021-04-300.119498
362021-05-310.119149
372021-06-300.118857
382021-07-310.119719
392021-08-310.119097
402021-09-300.118091
412021-10-310.124344
422021-11-300.120737
432021-12-310.116358
442022-01-310.114958
452022-02-280.114722
462022-03-310.113699
\n", "
" ], "text/plain": [ " ds yhat\n", "0 2018-05-31 0.124069\n", "1 2018-06-30 0.123136\n", "2 2018-07-31 0.124375\n", "3 2018-08-31 0.124707\n", "4 2018-09-30 0.124281\n", "5 2018-10-31 0.132125\n", "6 2018-11-30 0.127252\n", "7 2018-12-31 0.123648\n", "8 2019-01-31 0.121993\n", "9 2019-02-28 0.121721\n", "10 2019-03-31 0.122539\n", "11 2019-04-30 0.126377\n", "12 2019-05-31 0.126275\n", "13 2019-06-30 0.124292\n", "14 2019-07-31 0.125504\n", "15 2019-08-31 0.126429\n", "16 2019-09-30 0.126236\n", "17 2019-10-31 0.135017\n", "18 2019-11-30 0.127644\n", "19 2019-12-31 0.123715\n", "20 2020-01-31 0.120574\n", "21 2020-02-29 0.122055\n", "22 2020-03-31 0.117438\n", "23 2020-04-30 0.121737\n", "24 2020-05-31 0.121417\n", "25 2020-06-30 0.122117\n", "26 2020-07-31 0.123075\n", "27 2020-08-31 0.121882\n", "28 2020-09-30 0.120617\n", "29 2020-10-31 0.126160\n", "30 2020-11-30 0.123886\n", "31 2020-12-31 0.118808\n", "32 2021-01-31 0.117914\n", "33 2021-02-28 0.118066\n", "34 2021-03-31 0.115560\n", "35 2021-04-30 0.119498\n", "36 2021-05-31 0.119149\n", "37 2021-06-30 0.118857\n", "38 2021-07-31 0.119719\n", "39 2021-08-31 0.119097\n", "40 2021-09-30 0.118091\n", "41 2021-10-31 0.124344\n", "42 2021-11-30 0.120737\n", "43 2021-12-31 0.116358\n", "44 2022-01-31 0.114958\n", "45 2022-02-28 0.114722\n", "46 2022-03-31 0.113699" ] }, "execution_count": 48, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# yhat column is our prediction \n", "monthly_forecast[['ds','yhat']]" ] }, { "cell_type": "code", "execution_count": 49, "metadata": {}, "outputs": [], "source": [ "monthly_forecast[['ds','yhat']].to_csv('mx_over_42_prediction.csv')" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] } ], "metadata": { "kernelspec": { "display_name": "Python 3", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.7.5" } }, "nbformat": 4, "nbformat_minor": 2 }