{ "cells": [ { "cell_type": "code", "execution_count": 122, "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", "
BGGIdRatingUsername
02137888.0Tonydorrf
12137888.0tachyon14k
22137888.0Ungotter
32137888.0brainlocki3
42137888.0PPMP
............
189422101655213.0rseater
189422111655213.0Bluefox86
189422121655213.0serginator
189422131934881.0CaptainCattan
189422141934881.0trodat123
\n", "

18942215 rows × 3 columns

\n", "
" ], "text/plain": [ " BGGId Rating Username\n", "0 213788 8.0 Tonydorrf\n", "1 213788 8.0 tachyon14k\n", "2 213788 8.0 Ungotter\n", "3 213788 8.0 brainlocki3\n", "4 213788 8.0 PPMP\n", "... ... ... ...\n", "18942210 165521 3.0 rseater\n", "18942211 165521 3.0 Bluefox86\n", "18942212 165521 3.0 serginator\n", "18942213 193488 1.0 CaptainCattan\n", "18942214 193488 1.0 trodat123\n", "\n", "[18942215 rows x 3 columns]" ] }, "execution_count": 122, "metadata": {}, "output_type": "execute_result" } ], "source": [ "import pandas as pd\n", "from sklearn.preprocessing import LabelEncoder\n", "import numpy as np\n", "\n", "df = pd.read_csv(\"user_ratings.csv\")\n", "df" ] }, { "cell_type": "code", "execution_count": 123, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "array([195785, 385050, 200126, ..., 371648, 30181, 394601],\n", " shape=(18942215,))" ] }, "execution_count": 123, "metadata": {}, "output_type": "execute_result" } ], "source": [ "le = LabelEncoder()\n", "le.fit_transform(df['Username'])" ] }, { "cell_type": "code", "execution_count": 124, "metadata": {}, "outputs": [], "source": [ "df['Username_Encoded'] = le.transform(df['Username'])" ] }, { "cell_type": "code", "execution_count": 125, "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", "
BGGIdRatingUsernameUsername_Encoded
0213788.08.0Tonydorrf195785.0
103768206593.08.0Tonydorrf195785.0
147923267333.07.0Tonydorrf195785.0
171368249824.06.0Tonydorrf195785.0
184232301085.06.0Tonydorrf195785.0
...............
18801954252314.07.0Tonydorrf195785.0
18826671176247.08.0Tonydorrf195785.0
18901732259857.07.0Tonydorrf195785.0
18905081246530.07.0Tonydorrf195785.0
1891166989222.02.0Tonydorrf195785.0
\n", "

253 rows × 4 columns

\n", "
" ], "text/plain": [ " BGGId Rating Username Username_Encoded\n", "0 213788.0 8.0 Tonydorrf 195785.0\n", "103768 206593.0 8.0 Tonydorrf 195785.0\n", "147923 267333.0 7.0 Tonydorrf 195785.0\n", "171368 249824.0 6.0 Tonydorrf 195785.0\n", "184232 301085.0 6.0 Tonydorrf 195785.0\n", "... ... ... ... ...\n", "18801954 252314.0 7.0 Tonydorrf 195785.0\n", "18826671 176247.0 8.0 Tonydorrf 195785.0\n", "18901732 259857.0 7.0 Tonydorrf 195785.0\n", "18905081 246530.0 7.0 Tonydorrf 195785.0\n", "18911669 89222.0 2.0 Tonydorrf 195785.0\n", "\n", "[253 rows x 4 columns]" ] }, "execution_count": 125, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# Depcting that 'Tonydorrf' has a specific encoded value with multiple Board Game Reviews\n", "df.where(df['Username'] == 'Tonydorrf').dropna()" ] }, { "cell_type": "code", "execution_count": 126, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "BGGId int32\n", "Rating int16\n", "Username object\n", "Username_Encoded int32\n", "dtype: object" ] }, "execution_count": 126, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df['Username_Encoded'] = df['Username_Encoded'].astype(np.int32)\n", "df['Rating'] = df['Rating'].astype(np.int16)\n", "df['BGGId'] = df['BGGId'].astype(np.int32)\n", "df.dtypes" ] }, { "cell_type": "code", "execution_count": 127, "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", "
BGGIdRatingUsernameUsername_Encoded
02137888Tonydorrf195785
12137888tachyon14k385050
22137888Ungotter200126
32137888brainlocki3236689
42137888PPMP144112
...............
189422101655213rseater365759
189422111655213Bluefox8622977
189422121655213serginator371648
189422131934881CaptainCattan30181
189422141934881trodat123394601
\n", "

18942215 rows × 4 columns

\n", "
" ], "text/plain": [ " BGGId Rating Username Username_Encoded\n", "0 213788 8 Tonydorrf 195785\n", "1 213788 8 tachyon14k 385050\n", "2 213788 8 Ungotter 200126\n", "3 213788 8 brainlocki3 236689\n", "4 213788 8 PPMP 144112\n", "... ... ... ... ...\n", "18942210 165521 3 rseater 365759\n", "18942211 165521 3 Bluefox86 22977\n", "18942212 165521 3 serginator 371648\n", "18942213 193488 1 CaptainCattan 30181\n", "18942214 193488 1 trodat123 394601\n", "\n", "[18942215 rows x 4 columns]" ] }, "execution_count": 127, "metadata": {}, "output_type": "execute_result" } ], "source": [ "df" ] }, { "cell_type": "code", "execution_count": 128, "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "/var/folders/tw/7t5t3x_11sqflkfn0yxds1lr0000gq/T/ipykernel_4739/3299505638.py:2: DeprecationWarning: DataFrameGroupBy.apply operated on the grouping columns. This behavior is deprecated, and in a future version of pandas the grouping columns will be excluded from the operation. Either pass `include_groups=False` to exclude the groupings or explicitly select the grouping columns after groupby to silence this warning.\n", " df_sampled = df.groupby('BGGId').apply(lambda x: x.sample(frac=0.001, random_state=42)).reset_index(drop=True)\n" ] }, { "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", "
BGGIdRatingUsernameUsername_Encoded
019thedirtyhamm388576
117RocketSkunk161894
216Craigthulu38438
319Lassonius108428
417rainian44358660
...............
171153386287magnitt323191
171163404667Bayushi Sezaru18065
171173429429DoctorPhil50091
171183439058LanceCorporal107914
171193467038VG182200834
\n", "

17120 rows × 4 columns

\n", "
" ], "text/plain": [ " BGGId Rating Username Username_Encoded\n", "0 1 9 thedirtyhamm 388576\n", "1 1 7 RocketSkunk 161894\n", "2 1 6 Craigthulu 38438\n", "3 1 9 Lassonius 108428\n", "4 1 7 rainian44 358660\n", "... ... ... ... ...\n", "17115 338628 7 magnitt 323191\n", "17116 340466 7 Bayushi Sezaru 18065\n", "17117 342942 9 DoctorPhil 50091\n", "17118 343905 8 LanceCorporal 107914\n", "17119 346703 8 VG182 200834\n", "\n", "[17120 rows x 4 columns]" ] }, "execution_count": 128, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# Randomly select .1% of the data, in stretches of consecutive BGGId values\n", "df_sampled = df.groupby('BGGId').apply(lambda x: x.sample(frac=0.001, random_state=42)).reset_index(drop=True)\n", "\n", "#df_sampled = df.sample(frac=0.01, random_state=42)\n", "df_sampled" ] }, { "cell_type": "code", "execution_count": 129, "metadata": {}, "outputs": [], "source": [ "# Save the sampled data to a new CSV file\n", "df_sampled.to_csv(\"user_ratings_sampled.csv\", index=False)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# Create a recommendation System that:\n", "- Recommend *New* BoardGameIDs to a User\n", " - Inputs a user\n", " - Outpus BoardGame Recommendations\n", "- Allows a user to see which BoardGames are similar to another BoardGame\n", " - Inputs a BoardGame\n", " - Ouputs a Recommended BoardGame" ] }, { "cell_type": "code", "execution_count": 130, "metadata": {}, "outputs": [], "source": [ "sampled_df = pd.read_csv(\"user_ratings_sampled.csv\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Quickly inspect the data (EDA)" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Using the BGGId and Username_Encoded Features; Create the Sparse (Mostly NaN) User-Item Matrix via pivot table" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Perform SVD, then recostruct the User-Item Matrix" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## User - User Colaborative Filtering" ] }, { "cell_type": "code", "execution_count": 131, "metadata": {}, "outputs": [], "source": [ "# Compute cosine similarity between users\n", "\n", "\n", "# Convert back to a DataFrame for ease of use\n", "\n" ] }, { "cell_type": "code", "execution_count": 132, "metadata": {}, "outputs": [], "source": [ "# View your user similarity matrix as a DataFrame" ] }, { "cell_type": "code", "execution_count": 133, "metadata": {}, "outputs": [], "source": [ "# Save the user similarity matrix to csv\n" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Item - Item Colaborative Filtering" ] }, { "cell_type": "code", "execution_count": 134, "metadata": {}, "outputs": [], "source": [ "# Compute cosine similarity between items\n", "\n", "\n", "# Convert to a DataFrame for ease of use\n" ] }, { "cell_type": "code", "execution_count": 135, "metadata": {}, "outputs": [], "source": [ "# View your user similarity matrix as a DataFrame" ] }, { "cell_type": "code", "execution_count": 136, "metadata": {}, "outputs": [], "source": [ "# Save the user similarity matrix to csv" ] } ], "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": "3.13.2" } }, "nbformat": 4, "nbformat_minor": 2 }