{ "cells": [ { "cell_type": "code", "execution_count": 1, "id": "056e9ac4-0910-4a86-a85f-c3f95ddc4cb0", "metadata": { "tags": [] }, "outputs": [], "source": [ "import numpy as np\n", "import pandas as pd" ] }, { "cell_type": "code", "execution_count": 2, "id": "9d179665-48fa-4f45-b623-0e08ef2fbb27", "metadata": { "tags": [] }, "outputs": [], "source": [ "!pip install retentioneering" ] }, { "cell_type": "code", "execution_count": 3, "id": "f937e528-df49-4609-a1e5-a60e4561d2a7", "metadata": { "tags": [] }, "outputs": [], "source": [ "from retentioneering import datasets\n", "from retentioneering.eventstream import Eventstream\n", "stream = datasets.load_simple_shop()" ] }, { "cell_type": "markdown", "id": "2b1e46d2-b743-4842-998c-54e17ef2ff8c", "metadata": {}, "source": [ "## Basic example" ] }, { "cell_type": "code", "execution_count": 4, "id": "76830741-8a54-4ffa-b209-2a1477abef70", "metadata": { "tags": [] }, "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", "
 user_iduser_id_sharecountcount_sharesequence_typeuser_id_sample
Sequence      
path_end3 7511.003 7510.09other[921396444]
path_start3 7511.003 7510.09other[302653277]
catalog3 6110.9614 5180.36other[539857357]
main2 3850.645 6350.14other[104008042]
cart1 9240.512 8420.07other[314468142]
product21 4300.382 1720.05other[205492887]
delivery_choice1 3560.361 6860.04other[720173080]
product11 1220.301 5150.04other[704020228]
payment_choice9580.261 1070.03other[653107020]
delivery_courier7480.208340.02other[219678075]
payment_done6530.177060.02other[725600922]
payment_card5210.145650.01other[81533266]
delivery_pickup4690.135060.01other[747713930]
payment_cash1900.051970.00other[742899356]
\n" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/plain": [ "" ] }, "execution_count": 4, "metadata": {}, "output_type": "execute_result" } ], "source": [ "stream.sequences()" ] }, { "cell_type": "markdown", "id": "30f46bf4-ec3a-4ceb-9c77-8b64ebec9cb1", "metadata": {}, "source": [ "## Tuning the arguments" ] }, { "cell_type": "code", "execution_count": 5, "id": "574a611d-f45b-4963-8663-b10d90bbbd78", "metadata": { "tags": [] }, "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", "
 countcount_sharesession_id_sharesequence_typesession_id_sample
Sequence     
catalog -> catalog4 8570.060.40loop['736786060_1', '696093962_2', '226908074_1']
main -> catalog4 0640.050.51other['962262431_2', '206756025_2', '685067898_1']
session_start -> main3 7680.040.58other['392174907_3', '934489861_3', '536170195_3']
path_start -> session_start3 7510.040.58other['680256022_1', '110374635_1', '55520661_1']
session_end -> path_end3 7510.040.58other['256366749_2', '353256438_1', '540528050_2']
catalog -> session_end2 8520.030.44other['336287644_2', '623854494_2', '981878337_2']
path_start -> session_start -> catalog2 6860.030.42other['802760330_1', '236467681_1', '934489861_1']
session_start -> catalog2 6860.030.42other['161089315_1', '110335585_1', '686865201_1']
session_start -> main -> catalog2 6190.030.41other['835353025_1', '737390690_3', '716009752_2']
catalog -> product22 1720.030.27other['742899356_7', '709334840_1', '171887017_1']
catalog -> session_end -> path_end1 7200.020.27other['576372721_1', '70531100_2', '948950088_1']
catalog -> cart1 7090.020.24other['868120836_6', '3744822_2', '561151012_1']
cart -> delivery_choice1 6860.020.25other['402623898_1', '309507369_1', '112898732_2']
catalog -> catalog -> catalog1 6510.020.16loop['136612361_2', '708219612_1', '335829383_1']
catalog -> product11 5150.020.20other['800497846_2', '839611746_1', '794034682_5']
catalog -> main1 4130.020.18other['872575824_1', '827028867_1', '612188380_1']
main -> catalog -> catalog1 3610.020.20other['968662055_2', '157981126_6', '639143978_2']
catalog -> product2 -> catalog1 2980.020.16cycle['772579625_1', '343414552_1', '293620383_1']
product2 -> catalog1 2980.020.16other['221885049_1', '366154281_1', '740727955_1']
\n" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/plain": [ "" ] }, "execution_count": 5, "metadata": {}, "output_type": "execute_result" } ], "source": [ "stream\\\n", " .split_sessions(timeout=(30, 'm'))\\\n", " .sequences(\n", " ngram_range=(2, 3),\n", " weight_col='session_id',\n", " metrics=['count', 'count_share', 'paths_share'],\n", " threshold=['count', 1200],\n", " sorting=['count_share', False],\n", " heatmap_cols=['session_id_share'],\n", " sample_size=3\n", " )" ] }, { "cell_type": "markdown", "id": "05a0ccee-8879-4160-a827-7285dd683be0", "metadata": {}, "source": [ "## Comparing groups" ] }, { "cell_type": "code", "execution_count": 6, "id": "ff832330-8105-4d15-8b54-d069c1035f56", "metadata": { "tags": [] }, "outputs": [], "source": [ "np.random.seed(111)\n", "users = set(stream.to_dataframe()['user_id'])\n", "group1 = set(np.random.choice(list(users), size=len(users)//2))\n", "group2 = users - group1" ] }, { "cell_type": "code", "execution_count": 7, "id": "f549fae7-e844-46dc-9f6e-06c464ba0c9c", "metadata": { "tags": [] }, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "/Users/v.kukushkin/rete/retentioneering-tools-new-arch/retentioneering/tooling/sequences/sequences.py:134: FutureWarning: Passing a set as an indexer is deprecated and will raise in a future version. Use a list instead.\n", " vec_data.loc[groups[0], \"group\"] = group_names[0]\n", "/Users/v.kukushkin/rete/retentioneering-tools-new-arch/retentioneering/tooling/sequences/sequences.py:135: FutureWarning: Passing a set as an indexer is deprecated and will raise in a future version. Use a list instead.\n", " vec_data.loc[groups[1], \"group\"] = group_names[1]\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", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
 user_id_sharecount_sharesequence_typeuser_id_sample
 ABdelta_absdeltaABdelta_absdeltaAB
Sequence           
delivery_choice0.350.370.020.060.040.040.000.04other[799913972][333582245]
delivery_pickup0.120.130.000.020.010.010.000.03other[805683601][475256234]
cart0.510.520.010.020.070.070.000.03other[274110594][371912382]
payment_choice0.250.260.010.030.030.030.000.02other[258837206][358346667]
payment_done0.170.180.010.060.020.020.000.02other[900739025][962374537]
catalog0.960.970.010.010.360.370.010.01other[444244255][139276360]
product20.380.380.010.020.050.050.000.01other[573609710][407735919]
path_end1.001.000.000.000.090.090.000.00other[794982519][372695387]
path_start1.001.000.000.000.090.090.000.00other[410014458][613540564]
delivery_courier0.200.200.010.030.020.02-0.00-0.01other[110523244][902698857]
payment_card0.140.140.000.020.010.01-0.00-0.02other[490832159][238999718]
\n" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/plain": [ "" ] }, "execution_count": 7, "metadata": {}, "output_type": "execute_result" } ], "source": [ "stream.sequences(\n", " groups=[group1, group2],\n", " group_names=['A', 'B'],\n", " metrics=['paths_share', 'count_share'],\n", " threshold=[('user_id_share', 'delta_abs'), 0],\n", " sorting=[('count_share', 'delta'), False]\n", ")" ] } ], "metadata": { "kernelspec": { "display_name": "rete", "language": "python", "name": "rete" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.8.15" } }, "nbformat": 4, "nbformat_minor": 5 }