diff --git a/.ipynb_checkpoints/Untitled-checkpoint.ipynb b/.ipynb_checkpoints/Untitled-checkpoint.ipynb new file mode 100644 index 0000000..2fd6442 --- /dev/null +++ b/.ipynb_checkpoints/Untitled-checkpoint.ipynb @@ -0,0 +1,6 @@ +{ + "cells": [], + "metadata": {}, + "nbformat": 4, + "nbformat_minor": 2 +} diff --git a/.ipynb_checkpoints/Untitled1-checkpoint.ipynb b/.ipynb_checkpoints/Untitled1-checkpoint.ipynb new file mode 100644 index 0000000..027901f --- /dev/null +++ b/.ipynb_checkpoints/Untitled1-checkpoint.ipynb @@ -0,0 +1,393 @@ +{ + "cells": [ + { + "cell_type": "code", + "execution_count": 1, + "metadata": {}, + "outputs": [], + "source": [ + "import os\n", + "import sys\n", + "import matplotlib.pyplot as plt\n", + "import seaborn as sns\n", + "\n", + "nb_dir = os.path.split(os.getcwd())[0]\n", + "if nb_dir not in sys.path:\n", + " sys.path.append(nb_dir)\n", + " \n", + "from features.communication import *\n", + "import participants.query_db\n", + "\n", + "participants_inactive_usernames = participants.query_db.get_usernames()\n", + "df_sms = get_sms_data(participants_inactive_usernames)\n", + "df_calls = get_call_data(participants_inactive_usernames)" + ] + }, + { + "cell_type": "code", + "execution_count": 7, + "metadata": {}, + "outputs": [], + "source": [ + " count_calls = count_comms(df_calls)\n", + " count_sms = count_comms(df_sms)\n", + "\n", + " count_joined = count_calls.merge(\n", + " count_sms, on=\"participant_id\", suffixes=(\"_calls\", \"_sms\")\n", + " )" + ] + }, + { + "cell_type": "code", + "execution_count": 9, + "metadata": {}, + "outputs": [ + { + "data": { + "text/html": [ + "
\n", + " | participant_id | \n", + "no_incoming | \n", + "no_outgoing | \n", + "no_missed | \n", + "no_all_calls | \n", + "no_incoming_ratio | \n", + "no_outgoing_ratio | \n", + "duration_total_incoming | \n", + "duration_total_outgoing | \n", + "duration_max_incoming | \n", + "duration_max_outgoing | \n", + "no_received | \n", + "no_sent | \n", + "no_all_sms | \n", + "no_received_ratio | \n", + "no_sent_ratio | \n", + "
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
0 | \n", + "13 | \n", + "3.0 | \n", + "21.0 | \n", + "2.0 | \n", + "26.0 | \n", + "0.115385 | \n", + "0.807692 | \n", + "342.0 | \n", + "2836.0 | \n", + "196.0 | \n", + "355.0 | \n", + "7.0 | \n", + "7.0 | \n", + "14.0 | \n", + "0.500000 | \n", + "0.500000 | \n", + "
1 | \n", + "14 | \n", + "16.0 | \n", + "22.0 | \n", + "11.0 | \n", + "49.0 | \n", + "0.326531 | \n", + "0.448980 | \n", + "1873.0 | \n", + "2789.0 | \n", + "346.0 | \n", + "694.0 | \n", + "20.0 | \n", + "14.0 | \n", + "34.0 | \n", + "0.588235 | \n", + "0.411765 | \n", + "
2 | \n", + "15 | \n", + "3.0 | \n", + "2.0 | \n", + "NaN | \n", + "5.0 | \n", + "0.600000 | \n", + "0.400000 | \n", + "310.0 | \n", + "19.0 | \n", + "154.0 | \n", + "19.0 | \n", + "73.0 | \n", + "73.0 | \n", + "146.0 | \n", + "0.500000 | \n", + "0.500000 | \n", + "
3 | \n", + "16 | \n", + "4.0 | \n", + "6.0 | \n", + "3.0 | \n", + "13.0 | \n", + "0.307692 | \n", + "0.461538 | \n", + "1963.0 | \n", + "849.0 | \n", + "1037.0 | \n", + "638.0 | \n", + "8.0 | \n", + "2.0 | \n", + "10.0 | \n", + "0.800000 | \n", + "0.200000 | \n", + "
4 | \n", + "17 | \n", + "20.0 | \n", + "60.0 | \n", + "8.0 | \n", + "88.0 | \n", + "0.227273 | \n", + "0.681818 | \n", + "5789.0 | \n", + "17046.0 | \n", + "1966.0 | \n", + "3830.0 | \n", + "7.0 | \n", + "1.0 | \n", + "8.0 | \n", + "0.875000 | \n", + "0.125000 | \n", + "
... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "
56 | \n", + "91 | \n", + "15.0 | \n", + "13.0 | \n", + "3.0 | \n", + "31.0 | \n", + "0.483871 | \n", + "0.419355 | \n", + "3443.0 | \n", + "3636.0 | \n", + "644.0 | \n", + "1315.0 | \n", + "83.0 | \n", + "44.0 | \n", + "127.0 | \n", + "0.653543 | \n", + "0.346457 | \n", + "
57 | \n", + "92 | \n", + "3.0 | \n", + "4.0 | \n", + "1.0 | \n", + "8.0 | \n", + "0.375000 | \n", + "0.500000 | \n", + "231.0 | \n", + "648.0 | \n", + "167.0 | \n", + "433.0 | \n", + "4.0 | \n", + "6.0 | \n", + "10.0 | \n", + "0.400000 | \n", + "0.600000 | \n", + "
58 | \n", + "93 | \n", + "22.0 | \n", + "20.0 | \n", + "9.0 | \n", + "51.0 | \n", + "0.431373 | \n", + "0.392157 | \n", + "2534.0 | \n", + "1444.0 | \n", + "443.0 | \n", + "672.0 | \n", + "48.0 | \n", + "19.0 | \n", + "67.0 | \n", + "0.716418 | \n", + "0.283582 | \n", + "
59 | \n", + "106 | \n", + "12.0 | \n", + "30.0 | \n", + "6.0 | \n", + "48.0 | \n", + "0.250000 | \n", + "0.625000 | \n", + "3049.0 | \n", + "2637.0 | \n", + "878.0 | \n", + "380.0 | \n", + "10.0 | \n", + "10.0 | \n", + "20.0 | \n", + "0.500000 | \n", + "0.500000 | \n", + "
60 | \n", + "107 | \n", + "11.0 | \n", + "42.0 | \n", + "13.0 | \n", + "66.0 | \n", + "0.166667 | \n", + "0.636364 | \n", + "3804.0 | \n", + "9977.0 | \n", + "1519.0 | \n", + "1943.0 | \n", + "80.0 | \n", + "176.0 | \n", + "256.0 | \n", + "0.312500 | \n", + "0.687500 | \n", + "
61 rows × 16 columns
\n", + "\n", + " | id | \n", + "_id | \n", + "timestamp | \n", + "device_id | \n", + "call_type | \n", + "call_duration | \n", + "trace | \n", + "participant_id | \n", + "username | \n", + "freq | \n", + "contact_id | \n", + "
---|---|---|---|---|---|---|---|---|---|---|---|
0 | \n", + "118 | \n", + "1 | \n", + "1581936081010 | \n", + "78082f9f-98c2-468d-b4a2-7c835bd812bd | \n", + "2 | \n", + "0 | \n", + "87ae5eb2c5b7fe30bea2821e2ec052453d89ea6b | \n", + "13 | \n", + "uploader_20449 | \n", + "21 | \n", + "0 | \n", + "
1 | \n", + "176 | \n", + "3 | \n", + "1582295247982 | \n", + "78082f9f-98c2-468d-b4a2-7c835bd812bd | \n", + "1 | \n", + "196 | \n", + "87ae5eb2c5b7fe30bea2821e2ec052453d89ea6b | \n", + "13 | \n", + "uploader_20449 | \n", + "21 | \n", + "0 | \n", + "
2 | \n", + "209 | \n", + "4 | \n", + "1582305634014 | \n", + "78082f9f-98c2-468d-b4a2-7c835bd812bd | \n", + "2 | \n", + "237 | \n", + "87ae5eb2c5b7fe30bea2821e2ec052453d89ea6b | \n", + "13 | \n", + "uploader_20449 | \n", + "21 | \n", + "0 | \n", + "
3 | \n", + "210 | \n", + "5 | \n", + "1582561530334 | \n", + "78082f9f-98c2-468d-b4a2-7c835bd812bd | \n", + "2 | \n", + "126 | \n", + "87ae5eb2c5b7fe30bea2821e2ec052453d89ea6b | \n", + "13 | \n", + "uploader_20449 | \n", + "21 | \n", + "0 | \n", + "
4 | \n", + "253 | \n", + "6 | \n", + "1582627576077 | \n", + "78082f9f-98c2-468d-b4a2-7c835bd812bd | \n", + "2 | \n", + "255 | \n", + "87ae5eb2c5b7fe30bea2821e2ec052453d89ea6b | \n", + "13 | \n", + "uploader_20449 | \n", + "21 | \n", + "0 | \n", + "
... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "
4645 | \n", + "5874 | \n", + "131 | \n", + "1625602519843 | \n", + "4b62a655-cbf0-4ac0-a448-06726f45b56a | \n", + "2 | \n", + "2270 | \n", + "f4ebca8dc7305fe424d6bf7fbcd2e5086f98b453 | \n", + "90 | \n", + "uploader_53573 | \n", + "6 | \n", + "9 | \n", + "
4646 | \n", + "5882 | \n", + "139 | \n", + "1625753023456 | \n", + "4b62a655-cbf0-4ac0-a448-06726f45b56a | \n", + "3 | \n", + "0 | \n", + "2e5d63f6fddca2b66be810b5946c42eda24f2dbe | \n", + "90 | \n", + "uploader_53573 | \n", + "2 | \n", + "13 | \n", + "
4647 | \n", + "5883 | \n", + "140 | \n", + "1625754998767 | \n", + "4b62a655-cbf0-4ac0-a448-06726f45b56a | \n", + "3 | \n", + "0 | \n", + "2e5d63f6fddca2b66be810b5946c42eda24f2dbe | \n", + "90 | \n", + "uploader_53573 | \n", + "2 | \n", + "13 | \n", + "
4648 | \n", + "5884 | \n", + "141 | \n", + "1625823008392 | \n", + "4b62a655-cbf0-4ac0-a448-06726f45b56a | \n", + "2 | \n", + "0 | \n", + "7316d58b7bb7de097a2421c56010ac024a489451 | \n", + "90 | \n", + "uploader_53573 | \n", + "1 | \n", + "21 | \n", + "
4649 | \n", + "5903 | \n", + "158 | \n", + "1626110930233 | \n", + "4b62a655-cbf0-4ac0-a448-06726f45b56a | \n", + "2 | \n", + "53 | \n", + "7db4e9acf7c73837ddecdae5da523a28c774ba94 | \n", + "90 | \n", + "uploader_53573 | \n", + "1 | \n", + "24 | \n", + "
4650 rows × 11 columns
\n", + "\n", + " | id | \n", + "_id | \n", + "timestamp | \n", + "device_id | \n", + "call_type | \n", + "call_duration | \n", + "trace | \n", + "participant_id | \n", + "username | \n", + "freq | \n", + "contact_id | \n", + "total_call_duration | \n", + "no_contacts | \n", + "
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
0 | \n", + "118 | \n", + "1 | \n", + "1581936081010 | \n", + "78082f9f-98c2-468d-b4a2-7c835bd812bd | \n", + "2 | \n", + "0 | \n", + "87ae5eb2c5b7fe30bea2821e2ec052453d89ea6b | \n", + "13 | \n", + "uploader_20449 | \n", + "21 | \n", + "0 | \n", + "2844 | \n", + "5 | \n", + "
1 | \n", + "176 | \n", + "3 | \n", + "1582295247982 | \n", + "78082f9f-98c2-468d-b4a2-7c835bd812bd | \n", + "1 | \n", + "196 | \n", + "87ae5eb2c5b7fe30bea2821e2ec052453d89ea6b | \n", + "13 | \n", + "uploader_20449 | \n", + "21 | \n", + "0 | \n", + "2844 | \n", + "5 | \n", + "
2 | \n", + "209 | \n", + "4 | \n", + "1582305634014 | \n", + "78082f9f-98c2-468d-b4a2-7c835bd812bd | \n", + "2 | \n", + "237 | \n", + "87ae5eb2c5b7fe30bea2821e2ec052453d89ea6b | \n", + "13 | \n", + "uploader_20449 | \n", + "21 | \n", + "0 | \n", + "2844 | \n", + "5 | \n", + "
3 | \n", + "210 | \n", + "5 | \n", + "1582561530334 | \n", + "78082f9f-98c2-468d-b4a2-7c835bd812bd | \n", + "2 | \n", + "126 | \n", + "87ae5eb2c5b7fe30bea2821e2ec052453d89ea6b | \n", + "13 | \n", + "uploader_20449 | \n", + "21 | \n", + "0 | \n", + "2844 | \n", + "5 | \n", + "
4 | \n", + "253 | \n", + "6 | \n", + "1582627576077 | \n", + "78082f9f-98c2-468d-b4a2-7c835bd812bd | \n", + "2 | \n", + "255 | \n", + "87ae5eb2c5b7fe30bea2821e2ec052453d89ea6b | \n", + "13 | \n", + "uploader_20449 | \n", + "21 | \n", + "0 | \n", + "2844 | \n", + "5 | \n", + "
... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "
4352 | \n", + "5560 | \n", + "59 | \n", + "1621275689589 | \n", + "9f54e35c-d7cb-4f4c-8dc1-17dc86f2635e | \n", + "2 | \n", + "0 | \n", + "5eb72fe829c2af4a654007220119bdcf47499555 | \n", + "107 | \n", + "uploader_89606 | \n", + "1 | \n", + "17 | \n", + "0 | \n", + "22 | \n", + "
4336 | \n", + "5443 | \n", + "29 | \n", + "1620746142636 | \n", + "9f54e35c-d7cb-4f4c-8dc1-17dc86f2635e | \n", + "2 | \n", + "189 | \n", + "6dd761532337dfe596eb2e34f4c91216b38e28e2 | \n", + "107 | \n", + "uploader_89606 | \n", + "1 | \n", + "18 | \n", + "189 | \n", + "22 | \n", + "
4316 | \n", + "5237 | \n", + "10 | \n", + "1620140109908 | \n", + "9f54e35c-d7cb-4f4c-8dc1-17dc86f2635e | \n", + "2 | \n", + "85 | \n", + "9c4eab1dfc0114aecd64a7f594977acc9ab7936c | \n", + "107 | \n", + "uploader_89606 | \n", + "1 | \n", + "19 | \n", + "85 | \n", + "22 | \n", + "
4347 | \n", + "5524 | \n", + "44 | \n", + "1620971679122 | \n", + "9f54e35c-d7cb-4f4c-8dc1-17dc86f2635e | \n", + "1 | \n", + "120 | \n", + "a9fa73b6137d09288429de20172095978730e4b8 | \n", + "107 | \n", + "uploader_89606 | \n", + "1 | \n", + "20 | \n", + "120 | \n", + "22 | \n", + "
4331 | \n", + "5364 | \n", + "26 | \n", + "1620630328635 | \n", + "9f54e35c-d7cb-4f4c-8dc1-17dc86f2635e | \n", + "2 | \n", + "184 | \n", + "cfe98eee4a27b377f4cde1ea5c39d24d0475b533 | \n", + "107 | \n", + "uploader_89606 | \n", + "1 | \n", + "21 | \n", + "184 | \n", + "22 | \n", + "
4650 rows × 13 columns
\n", + "\n", + " | id | \n", + "_id | \n", + "timestamp | \n", + "device_id | \n", + "message_type | \n", + "trace | \n", + "participant_id | \n", + "username | \n", + "freq | \n", + "contact_id | \n", + "no_contacts | \n", + "
---|---|---|---|---|---|---|---|---|---|---|---|
195 | \n", + "379 | \n", + "7 | \n", + "1582964434597 | \n", + "78082f9f-98c2-468d-b4a2-7c835bd812bd | \n", + "2 | \n", + "417b9c87f5b573530bcffba8577777b3a964d671 | \n", + "13 | \n", + "uploader_20449 | \n", + "5 | \n", + "0 | \n", + "6 | \n", + "
196 | \n", + "380 | \n", + "8 | \n", + "1582964434974 | \n", + "78082f9f-98c2-468d-b4a2-7c835bd812bd | \n", + "2 | \n", + "417b9c87f5b573530bcffba8577777b3a964d671 | \n", + "13 | \n", + "uploader_20449 | \n", + "5 | \n", + "0 | \n", + "6 | \n", + "
197 | \n", + "382 | \n", + "10 | \n", + "1582965988609 | \n", + "78082f9f-98c2-468d-b4a2-7c835bd812bd | \n", + "2 | \n", + "417b9c87f5b573530bcffba8577777b3a964d671 | \n", + "13 | \n", + "uploader_20449 | \n", + "5 | \n", + "0 | \n", + "6 | \n", + "
198 | \n", + "383 | \n", + "11 | \n", + "1582965988873 | \n", + "78082f9f-98c2-468d-b4a2-7c835bd812bd | \n", + "2 | \n", + "417b9c87f5b573530bcffba8577777b3a964d671 | \n", + "13 | \n", + "uploader_20449 | \n", + "5 | \n", + "0 | \n", + "6 | \n", + "
199 | \n", + "396 | \n", + "12 | \n", + "1582965988873 | \n", + "78082f9f-98c2-468d-b4a2-7c835bd812bd | \n", + "2 | \n", + "417b9c87f5b573530bcffba8577777b3a964d671 | \n", + "13 | \n", + "uploader_20449 | \n", + "5 | \n", + "0 | \n", + "6 | \n", + "
... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "
5193 | \n", + "6137 | \n", + "4 | \n", + "1619789360665 | \n", + "9f54e35c-d7cb-4f4c-8dc1-17dc86f2635e | \n", + "2 | \n", + "2340c1d2b9e5d550373423a599014468a4dc3678 | \n", + "107 | \n", + "uploader_89606 | \n", + "3 | \n", + "12 | \n", + "16 | \n", + "
5194 | \n", + "6135 | \n", + "2 | \n", + "1619787273829 | \n", + "9f54e35c-d7cb-4f4c-8dc1-17dc86f2635e | \n", + "1 | \n", + "2340c1d2b9e5d550373423a599014468a4dc3678 | \n", + "107 | \n", + "uploader_89606 | \n", + "3 | \n", + "12 | \n", + "16 | \n", + "
5417 | \n", + "6690 | \n", + "210 | \n", + "1620980437198 | \n", + "9f54e35c-d7cb-4f4c-8dc1-17dc86f2635e | \n", + "1 | \n", + "198d7e63a2b4d8a7ca9bb92f74e6974ca17edc56 | \n", + "107 | \n", + "uploader_89606 | \n", + "1 | \n", + "13 | \n", + "16 | \n", + "
5447 | \n", + "6770 | \n", + "256 | \n", + "1621407668019 | \n", + "9f54e35c-d7cb-4f4c-8dc1-17dc86f2635e | \n", + "1 | \n", + "d4a67b53e704247de47064850efd3647e8dcaffb | \n", + "107 | \n", + "uploader_89606 | \n", + "1 | \n", + "14 | \n", + "16 | \n", + "
5440 | \n", + "6742 | \n", + "248 | \n", + "1621253313544 | \n", + "9f54e35c-d7cb-4f4c-8dc1-17dc86f2635e | \n", + "1 | \n", + "ffbc6a5f0f601cf2d9cdad9d3a588633e1a1967c | \n", + "107 | \n", + "uploader_89606 | \n", + "1 | \n", + "15 | \n", + "16 | \n", + "
5864 rows × 11 columns
\n", + "\n", + " | participant_id | \n", + "no_calls_no_sms_ratio | \n", + "no_incoming_calls_no_recieved_sms_ratio | \n", + "no_outgoing_calls_no_sent_sms_ratio | \n", + "no_calls_contacts_no_sms_contacts_ratio | \n", + "
---|---|---|---|---|---|
0 | \n", + "13 | \n", + "0.650000 | \n", + "0.700000 | \n", + "0.750000 | \n", + "0.454545 | \n", + "
1 | \n", + "14 | \n", + "0.590361 | \n", + "0.555556 | \n", + "0.611111 | \n", + "0.714286 | \n", + "
2 | \n", + "15 | \n", + "0.033113 | \n", + "0.960526 | \n", + "0.026667 | \n", + "0.173913 | \n", + "
3 | \n", + "16 | \n", + "0.565217 | \n", + "0.666667 | \n", + "0.750000 | \n", + "0.666667 | \n", + "
4 | \n", + "17 | \n", + "0.916667 | \n", + "0.259259 | \n", + "0.983607 | \n", + "0.857143 | \n", + "
... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "... | \n", + "
56 | \n", + "91 | \n", + "0.196203 | \n", + "0.846939 | \n", + "0.228070 | \n", + "0.666667 | \n", + "
57 | \n", + "92 | \n", + "0.444444 | \n", + "0.571429 | \n", + "0.400000 | \n", + "0.600000 | \n", + "
58 | \n", + "93 | \n", + "0.432203 | \n", + "0.685714 | \n", + "0.512821 | \n", + "0.428571 | \n", + "
59 | \n", + "106 | \n", + "0.705882 | \n", + "0.454545 | \n", + "0.750000 | \n", + "0.769231 | \n", + "
60 | \n", + "107 | \n", + "0.204969 | \n", + "0.879121 | \n", + "0.192661 | \n", + "0.578947 | \n", + "
61 rows × 5 columns
\n", + "