matti_jms_collabs/preparedness/main.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "1816c783-1fbb-4f6a-8465-2504face5756",
   "metadata": {},
   "source": [
    "# Visualize Most Prepared Users\n",
    "This workflow loads user embeddings, generates a prompt embedding, computes similarity, and visualizes the most prepared users."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "46d59443-3ca9-451a-8132-819c891817fa",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/tqdm/auto.py:21: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
      "  from .autonotebook import tqdm as notebook_tqdm\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/numba/np/ufunc/dufunc.py:344: NumbaWarning: Compilation requested for previously compiled argument types ((uint32,)). This has no effect and perhaps indicates a bug in the calling code (compiling a ufunc more than once for the same signature\n",
      "  warnings.warn(msg, errors.NumbaWarning)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/numba/np/ufunc/dufunc.py:344: NumbaWarning: Compilation requested for previously compiled argument types ((uint32,)). This has no effect and perhaps indicates a bug in the calling code (compiling a ufunc more than once for the same signature\n",
      "  warnings.warn(msg, errors.NumbaWarning)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/numba/np/ufunc/dufunc.py:344: NumbaWarning: Compilation requested for previously compiled argument types ((uint32,)). This has no effect and perhaps indicates a bug in the calling code (compiling a ufunc more than once for the same signature\n",
      "  warnings.warn(msg, errors.NumbaWarning)\n"
     ]
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "from sklearn.metrics.pairwise import cosine_similarity\n",
    "from sklearn.decomposition import PCA\n",
    "from sklearn.manifold import TSNE\n",
    "import plotly.express as px\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import umap\n",
    "import umap.plot\n",
    "from utils import get_ollama_embedding"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "03e6e9fb-49cc-4127-af4f-d874ea43c3aa",
   "metadata": {},
   "source": [
    " ## Read in the file with user vectors\n",
    " Update the path/format as needed."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c2064964-73d6-4e44-9dea-84f00e61b8ef",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "<ipython-input-2-7166d4ff9397>:3: DtypeWarning: Columns (58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84) have mixed types. Specify dtype option on import or set low_memory=False.\n",
      "  pd.read_csv(\"./data/eurobarometer_preparedness_model_data_v3.csv\")\n"
     ]
    }
   ],
   "source": [
    "df = (\n",
    "    pd.read_csv(\"./data/eurobarometer_preparedness_model_data_v3.csv\")\n",
    "    .reset_index()\n",
    "    .assign(\n",
    "        **{\n",
    "            \"user_id\": lambda x: x[\"index\"].astype(str)\n",
    "            + \"_\"\n",
    "            + x[\"country_code_iso_3166\"]\n",
    "            + \"_\"\n",
    "            + x[\"age_recoded_6_categories\"].astype(str)\n",
    "            + \"_\"\n",
    "            + x[\"gender\"].astype(str)\n",
    "        }\n",
    "    )\n",
    "    # select only relevant countries for visualization\n",
    "    .loc[\n",
    "        lambda x: x[\"country_code_iso_3166\"].isin(\n",
    "            [\"FI\", \"DE-E\", \"DE-W\", \"FR\", \"ES\", \"PT\"]\n",
    "        )\n",
    "    ]\n",
    ")\n",
    "df.shape\n",
    "\n",
    "# relabel None of the above/ Non binary/ do not recognize yourself in above categories/Prefer not to say to other\n",
    "df[\"gender\"] = df[\"gender\"].replace({\n",
    "    \"None of the above/ Non binary/ do not recognize yourself in above categories/Prefer not to say\": \"Other\",\n",
    "})"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2dd1f1f6-5113-41b0-b693-458fe455db53",
   "metadata": {},
   "source": [
    " ## Convert string embeddings to lists if needed"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "16003dff-9e14-43e3-8bd7-26b9a245a8bf",
   "metadata": {},
   "outputs": [],
   "source": [
    "def parse_embedding(x):\n",
    "    if isinstance(x, str):\n",
    "        return [float(i) for i in x.strip(\"[]\").split(\",\")]\n",
    "    return x\n",
    "\n",
    "\n",
    "df[\"embedding\"] = df[\"embedding\"].apply(parse_embedding)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c734672e-f68d-4688-b7f4-b1f8e56859cd",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of best users: 2\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[{'user_id': '7977_ES_35-44_Woman',\n",
       "  'user_text': 'risks_cntry_most_exposed_to_firstly: Terrorist attacks | risks_pers_most_exposed_to_firstly: Extreme weather events (violent storms, droughts, heatwaves, cold waves, etc.) | risks_pers_most_exposed_to_number_of_mentioned_risks: 2 mentions | pot_info_sources_to_learn_about_disaster_risks_firstly: National media | statements_disaster_risks_readseenheard_info_in_last_12m: Tend to agree | statements_disaster_risks_feel_well_informed: Tend to agree | statements_disaster_risks_trust_information_by_pub_auth_on_risks_where_you_live: Totally agree | statements_disaster_risks_easy_to_find_information_by_pub_auth_on_risks_where_you_live: Tend to agree | statements_disaster_risks_know_where_to_find_info_when_travelling_to_oth_eu_cntry: Totally agree | disaster_measures_in_hh_emergency_supply_drinksfood: Keep an emergency supply stock/pack of drinks, food | disaster_measures_in_hh_emergency_supply_water_cookinghygiene: Keep an emergency supply of water for cooking and hygiene | disaster_measures_in_hh_flashlightcandles: Have flashlight or candles accessible | disaster_measures_in_hh_batterypowered_radio: Have a battery-powered radio accessible | disaster_measures_in_hh_emergency_pharmacy: Keep a home pharmacy for emergencies | disaster_measures_in_hh_copies_imp_documentsstored_safely: Have made sure you have copies of your most important documents or have stored them safely | disaster_measures_in_hh_emergency_grabbag: Have prepared a grab-bag, in case you need to evacuate rapidly in an emergency | disaster_measures_in_hh_signed_up_for_alerts: Have signed up for alerts and warnings from emergency services or authorities | disaster_measures_in_hh_participated_in_trainingexercise: Have participated in a training or exercise, to learn how to react in an emergency | disaster_measures_in_hh_informed_about_official_response_plan: Got informed on the response plan your city, region or country has for a disaster or emergency (e.g. (...) | disaster_measures_in_hh_agreed_with_friendsfamily_to_contact: Agreed with family, friends on how to contact each other in case of an emergency | disaster_measures_in_hh_discussed_common_prot_measures_in_neighbourhood: Discussed common protective measures in your neighbourhood | disaster_measures_in_hh_invested_in_prot_measures_in_home: Have invested in protective measures in your home (e.g. flood-proofed the electricity installation, cleared (...) | how_many_days_meet_water_needs_if_water_services_disrupted: More than 7 days | how_many_days_power_essent_appliances_if_elec_interrupted: More than 7 days | how_many_days_cook_mealsheat_if_gas_disrupted: More than 7 days | how_many_days_provide_food_if_transportation_disrupted: More than 7 days | how_many_days_continued_treatment_if_medication_supply_disrupted: More than 7 days | personal_disaster_preparedness_better_able_to_cope_by_prep: Totally agree | personal_disaster_preparedness_feel_well_prepared: Tend to agree | personal_disaster_preparedness_no_timefin_resources_to_prep: Tend to disagree | personal_disaster_preparedness_easy_to_find_info_on_how_to_prep: Tend to agree | personal_disaster_preparedness_need_more_info_to_prep: Totally agree | personal_disaster_preparedness_know_how_emerg_services_will_alert: Tend to agree | personal_disaster_preparedness_know_what_to_do_in_event_of_disaster: Tend to agree | personal_disaster_preparedness_employerschool_encourages_trainingprep: Totally disagree | personal_disaster_preparedness_emerg_services_encourage_trainingprep: Tend to agree | relying_on_in_first_days_of_disaster_familyfriends: A great deal | relying_on_in_first_days_of_disaster_people_in_neighbourhood: A great deal | relying_on_in_first_days_of_disaster_assocsnonprofit_orgs: A great deal | relying_on_in_first_days_of_disaster_emerg_services: A great deal | relying_on_in_first_days_of_disaster_local_authgvmt_services: A great deal | relying_on_in_first_days_of_disaster_workemployerschooledu_institution: Not a lot | relying_on_in_first_days_of_disaster_private_sector_entities: Not a lot | trust_in_emerg_services_to_handle_disastersemerg_situations_properly: Tend to trust | engaging_in_voluntary_work_for_emerg_responder_orgs: No, you have never engaged in voluntary work and do not plan to do so | occupation_of_respondent: Skilled manual worker | age_recoded_6_categories: 35-44 | size_of_community: Towns/suburbs | direction_things_are_going_life_personally: Things are going in the right direction | political_discussion_local_matters: Never | political_discussion_national_matters: Never | internet_use_total: Everyday/almost everyday (at least once 1 in d62_1 to d62_4) | gender: Woman | age_education: 23.0 | standard_of_living_last_5yrs_in_light_of_crises: Your standard of living has not changed | personal_living_conditions_in_one_years_time: Worse | standard_of_living_next_5yrs: Your standard of living will not change | region_austria: nan | region_belgium: nan | region_bulgaria: nan | region_croatia: nan | region_cyprus: nan | region_czechia: nan | region_denmark: nan | region_germany: nan | region_estonia: nan | region_finland: nan | region_france: nan | region_greece: nan | region_hungary: nan | region_ireland: nan | region_italy: nan | region_latvia: nan | region_lithuania: nan | region_luxembourg: nan | region_malta: nan | region_netherlands: nan | region_poland: nan | region_portugal: nan | region_romania: nan | region_slovenia: nan | region_slovakia: nan | region_spain: ES61 - Andalucia | region_sweden: nan | education_level_preprimary_education_incl_no_education: Not mentioned | education_level_primary_education: Not mentioned | education_level_lower_secondary_education: Not mentioned | education_level_upper_secondary_education: Not mentioned | education_level_postsecondary_non_tertiary_incl_prevocationalvocational: Not mentioned | education_level_shortcycle_tertiary: Not mentioned | education_level_bachelor_or_equivalent: Bachelor or equivalent | education_level_master_or_equivalent: Not mentioned | education_level_doctoral_or_equivalent: Not mentioned'},\n",
       " {'user_id': '9221_FI_55-64_Man',\n",
       "  'user_text': 'risks_cntry_most_exposed_to_firstly: Cybersecurity threats (e.g. cyberattacks, cybercrimes, etc.) | risks_pers_most_exposed_to_firstly: Extreme weather events (violent storms, droughts, heatwaves, cold waves, etc.) | risks_pers_most_exposed_to_number_of_mentioned_risks: 3 mentions | pot_info_sources_to_learn_about_disaster_risks_firstly: Civil society organisations, non-profit organisations | statements_disaster_risks_readseenheard_info_in_last_12m: Tend to agree | statements_disaster_risks_feel_well_informed: Totally agree | statements_disaster_risks_trust_information_by_pub_auth_on_risks_where_you_live: Totally agree | statements_disaster_risks_easy_to_find_information_by_pub_auth_on_risks_where_you_live: Totally agree | statements_disaster_risks_know_where_to_find_info_when_travelling_to_oth_eu_cntry: Totally agree | disaster_measures_in_hh_emergency_supply_drinksfood: Keep an emergency supply stock/pack of drinks, food | disaster_measures_in_hh_emergency_supply_water_cookinghygiene: Keep an emergency supply of water for cooking and hygiene | disaster_measures_in_hh_flashlightcandles: Have flashlight or candles accessible | disaster_measures_in_hh_batterypowered_radio: Have a battery-powered radio accessible | disaster_measures_in_hh_emergency_pharmacy: Keep a home pharmacy for emergencies | disaster_measures_in_hh_copies_imp_documentsstored_safely: Have made sure you have copies of your most important documents or have stored them safely | disaster_measures_in_hh_emergency_grabbag: Have prepared a grab-bag, in case you need to evacuate rapidly in an emergency | disaster_measures_in_hh_signed_up_for_alerts: Have signed up for alerts and warnings from emergency services or authorities | disaster_measures_in_hh_participated_in_trainingexercise: Have participated in a training or exercise, to learn how to react in an emergency | disaster_measures_in_hh_informed_about_official_response_plan: Got informed on the response plan your city, region or country has for a disaster or emergency (e.g. (...) | disaster_measures_in_hh_agreed_with_friendsfamily_to_contact: Agreed with family, friends on how to contact each other in case of an emergency | disaster_measures_in_hh_discussed_common_prot_measures_in_neighbourhood: Discussed common protective measures in your neighbourhood | disaster_measures_in_hh_invested_in_prot_measures_in_home: Have invested in protective measures in your home (e.g. flood-proofed the electricity installation, cleared (...) | how_many_days_meet_water_needs_if_water_services_disrupted: More than 7 days | how_many_days_power_essent_appliances_if_elec_interrupted: 1 day or less | how_many_days_cook_mealsheat_if_gas_disrupted: More than 7 days | how_many_days_provide_food_if_transportation_disrupted: More than 7 days | how_many_days_continued_treatment_if_medication_supply_disrupted: More than 7 days | personal_disaster_preparedness_better_able_to_cope_by_prep: Totally agree | personal_disaster_preparedness_feel_well_prepared: Totally agree | personal_disaster_preparedness_no_timefin_resources_to_prep: Totally disagree | personal_disaster_preparedness_easy_to_find_info_on_how_to_prep: Totally agree | personal_disaster_preparedness_need_more_info_to_prep: Tend to disagree | personal_disaster_preparedness_know_how_emerg_services_will_alert: Totally agree | personal_disaster_preparedness_know_what_to_do_in_event_of_disaster: Totally agree | personal_disaster_preparedness_employerschool_encourages_trainingprep: Totally agree | personal_disaster_preparedness_emerg_services_encourage_trainingprep: Totally agree | relying_on_in_first_days_of_disaster_familyfriends: A great deal | relying_on_in_first_days_of_disaster_people_in_neighbourhood: Somewhat | relying_on_in_first_days_of_disaster_assocsnonprofit_orgs: Somewhat | relying_on_in_first_days_of_disaster_emerg_services: A great deal | relying_on_in_first_days_of_disaster_local_authgvmt_services: A great deal | relying_on_in_first_days_of_disaster_workemployerschooledu_institution: A great deal | relying_on_in_first_days_of_disaster_private_sector_entities: A great deal | trust_in_emerg_services_to_handle_disastersemerg_situations_properly: Totally trust | engaging_in_voluntary_work_for_emerg_responder_orgs: Yes, you are already currently engaged in voluntary work | occupation_of_respondent: Unemployed, temporarily not working | age_recoded_6_categories: 55-64 | size_of_community: Rural | direction_things_are_going_life_personally: Things are going in the right direction | political_discussion_local_matters: Occasionally | political_discussion_national_matters: Occasionally | internet_use_total: Everyday/almost everyday (at least once 1 in d62_1 to d62_4) | gender: Man | age_education: 16.0 | standard_of_living_last_5yrs_in_light_of_crises: Your standard of living has not changed | personal_living_conditions_in_one_years_time: The same | standard_of_living_next_5yrs: Your standard of living will not change | region_austria: nan | region_belgium: nan | region_bulgaria: nan | region_croatia: nan | region_cyprus: nan | region_czechia: nan | region_denmark: nan | region_germany: nan | region_estonia: nan | region_finland: FI1D - Pohjois- ja Itae-Suomi | region_france: nan | region_greece: nan | region_hungary: nan | region_ireland: nan | region_italy: nan | region_latvia: nan | region_lithuania: nan | region_luxembourg: nan | region_malta: nan | region_netherlands: nan | region_poland: nan | region_portugal: nan | region_romania: nan | region_slovenia: nan | region_slovakia: nan | region_spain: nan | region_sweden: nan | education_level_preprimary_education_incl_no_education: Not mentioned | education_level_primary_education: Not mentioned | education_level_lower_secondary_education: Not mentioned | education_level_upper_secondary_education: Not mentioned | education_level_postsecondary_non_tertiary_incl_prevocationalvocational: Post-secondary non tertiary (including pre-vocational or vocational education) | education_level_shortcycle_tertiary: Not mentioned | education_level_bachelor_or_equivalent: Not mentioned | education_level_master_or_equivalent: Not mentioned | education_level_doctoral_or_equivalent: Not mentioned'}]"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Select best user:\n",
    "# disaster_measures_in_hh_* == 1, and how_many_days_meet_* == (4 | 5)\n",
    "# Following relabeling in data_preparation_raw.py:\n",
    "# disaster_measures_in_hh_* != \"Not mentioned\"\n",
    "# how_many_days_meet_* == \"More than 7 days\"\n",
    "disaster_measures_in_hh_columns = [\n",
    "    col for col in df.columns if col.startswith(\"disaster_measures_in_hh_\")\n",
    "]\n",
    "how_many_days_meet_columns = [\n",
    "    col for col in df.columns if col.startswith(\"how_many_days_meet_\")\n",
    "]\n",
    "best_users = df[\n",
    "    (df[disaster_measures_in_hh_columns] != \"Not mentioned\").all(axis=1)\n",
    "    & (df[how_many_days_meet_columns] == \"More than 7 days\").all(axis=1)\n",
    "]\n",
    "print(f\"Number of best users: {len(best_users)}\")\n",
    "\n",
    "best_users.head(5)[[\"user_id\", \"user_text\"]].to_dict(orient=\"records\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0a447820-3ac9-4861-9e01-b1161eed9a7d",
   "metadata": {},
   "source": [
    " ## Write your prompt and generate its embedding"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6f78b8b8-c996-4c3f-8b7c-5c58865b6cb2",
   "metadata": {},
   "outputs": [],
   "source": [
    "# prompt = \"The user is highly prepared for disasters, with emergency supplies and a clear plan.\"\n",
    "prompt = best_users.head(1)[\"user_text\"].values[\n",
    "    0\n",
    "]  # Example: use the first user's text as the prompt\n",
    "prompt_embedding = get_ollama_embedding(prompt)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "695476dc-dd76-48b8-893f-ad63b45553ff",
   "metadata": {},
   "source": [
    " ## Compute similarity between each user and the prompt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bf83f4a4-2a41-4f8a-8d88-9efa0fba7aff",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:203: RuntimeWarning: divide by zero encountered in matmul\n",
      "  ret = a @ b\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:203: RuntimeWarning: overflow encountered in matmul\n",
      "  ret = a @ b\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:203: RuntimeWarning: invalid value encountered in matmul\n",
      "  ret = a @ b\n"
     ]
    }
   ],
   "source": [
    "user_embeddings = np.vstack(df[\"embedding\"].values)\n",
    "prompt_vec = np.array(prompt_embedding).reshape(1, -1)\n",
    "similarities = cosine_similarity(user_embeddings, prompt_vec).flatten()\n",
    "df[\"similarity\"] = similarities"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "12feab6d-70fa-4be7-9982-276af073c4af",
   "metadata": {},
   "source": [
    " ## Visualize the users by similarity"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "30bb12e5-b265-404a-a4a1-46f97fcd8f2f",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: divide by zero encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: overflow encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: invalid value encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: divide by zero encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: overflow encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: invalid value encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: divide by zero encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: overflow encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: invalid value encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: divide by zero encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: overflow encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: invalid value encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: divide by zero encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: overflow encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: invalid value encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: divide by zero encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: overflow encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: invalid value encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: divide by zero encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: overflow encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: invalid value encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: divide by zero encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: overflow encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: invalid value encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: divide by zero encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: overflow encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: invalid value encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: divide by zero encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: overflow encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/scipy/_lib/_util.py:1279: RuntimeWarning: invalid value encountered in vecdot\n",
      "  return np.vecdot(x1, x2, axis=axis)\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 500x700 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "_ = plt.figure(figsize=(5, 7))\n",
    "_ = sns.violinplot(\n",
    "    x=\"similarity\", y=\"country_code_iso_3166\", hue=\"country_code_iso_3166\", data=df\n",
    ")\n",
    "_ = sns.stripplot(\n",
    "    x=\"similarity\",\n",
    "    y=\"country_code_iso_3166\",\n",
    "    data=df,\n",
    "    hue=\"country_code_iso_3166\",\n",
    "    alpha=0.8,\n",
    "    jitter=True,\n",
    "    linewidth=0.5,\n",
    "    edgecolor=\"white\",\n",
    ")\n",
    "_ = plt.title(\"Distribution of User Similarities to Preparedness Prompt\")\n",
    "_ = plt.xlabel(\"Cosine Similarity\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b2c05876-7fd1-49bd-b494-202d45013e42",
   "metadata": {},
   "outputs": [],
   "source": [
    "# np.savetxt(\"./data/embeddings.tsv\", np.vstack(df[\"embedding\"].values), delimiter=\"\\t\")\n",
    "\n",
    "# # %%\n",
    "# df[[\"user_id\"]].to_csv(\n",
    "#     \"./data/metadata.tsv\", sep=\"\\t\", index=False\n",
    "# )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b5359221-39f4-4d6e-a958-37d732628615",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:350: RuntimeWarning: divide by zero encountered in matmul\n",
      "  Q, _ = normalizer(A @ Q)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:350: RuntimeWarning: overflow encountered in matmul\n",
      "  Q, _ = normalizer(A @ Q)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:350: RuntimeWarning: invalid value encountered in matmul\n",
      "  Q, _ = normalizer(A @ Q)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:351: RuntimeWarning: divide by zero encountered in matmul\n",
      "  Q, _ = normalizer(A.T @ Q)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:351: RuntimeWarning: overflow encountered in matmul\n",
      "  Q, _ = normalizer(A.T @ Q)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:351: RuntimeWarning: invalid value encountered in matmul\n",
      "  Q, _ = normalizer(A.T @ Q)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:355: RuntimeWarning: divide by zero encountered in matmul\n",
      "  Q, _ = qr_normalizer(A @ Q)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:355: RuntimeWarning: overflow encountered in matmul\n",
      "  Q, _ = qr_normalizer(A @ Q)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:355: RuntimeWarning: invalid value encountered in matmul\n",
      "  Q, _ = qr_normalizer(A @ Q)\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:577: RuntimeWarning: divide by zero encountered in matmul\n",
      "  B = Q.T @ M\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:577: RuntimeWarning: overflow encountered in matmul\n",
      "  B = Q.T @ M\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:577: RuntimeWarning: invalid value encountered in matmul\n",
      "  B = Q.T @ M\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:590: RuntimeWarning: divide by zero encountered in matmul\n",
      "  U = Q @ Uhat\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:590: RuntimeWarning: overflow encountered in matmul\n",
      "  U = Q @ Uhat\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:590: RuntimeWarning: invalid value encountered in matmul\n",
      "  U = Q @ Uhat\n"
     ]
    }
   ],
   "source": [
    "# Reduce embeddings to 3D with PCA\n",
    "pca = PCA(n_components=3)\n",
    "embeddings_3d = pca.fit_transform(np.vstack(df[\"embedding\"].values))\n",
    "\n",
    "# Add PCA components to dataframe\n",
    "df[\"pca1\"] = embeddings_3d[:, 0]\n",
    "df[\"pca2\"] = embeddings_3d[:, 1]\n",
    "df[\"pca3\"] = embeddings_3d[:, 2]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "634e62e8-5c95-4914-a964-08fc71374d42",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Interactive 3D scatter plot\n",
    "fig = px.scatter_3d(\n",
    "    df,\n",
    "    x=\"pca1\",\n",
    "    y=\"pca2\",\n",
    "    z=\"pca3\",\n",
    "    color=\"age_recoded_6_categories\",\n",
    "    hover_data=[\"user_id\", \"gender\", \"age_recoded_6_categories\"],\n",
    "    title=\"User Embeddings (PCA 3D)\"\n",
    ")\n",
    "fig.write_html(\"./output/user_embeddings_pca_3d.html\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5d459c9a-d452-4be5-b84f-72ccdff0fdac",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:350: RuntimeWarning:\n",
      "\n",
      "divide by zero encountered in matmul\n",
      "\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:350: RuntimeWarning:\n",
      "\n",
      "overflow encountered in matmul\n",
      "\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:350: RuntimeWarning:\n",
      "\n",
      "invalid value encountered in matmul\n",
      "\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:351: RuntimeWarning:\n",
      "\n",
      "divide by zero encountered in matmul\n",
      "\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:351: RuntimeWarning:\n",
      "\n",
      "overflow encountered in matmul\n",
      "\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:351: RuntimeWarning:\n",
      "\n",
      "invalid value encountered in matmul\n",
      "\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:355: RuntimeWarning:\n",
      "\n",
      "divide by zero encountered in matmul\n",
      "\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:355: RuntimeWarning:\n",
      "\n",
      "overflow encountered in matmul\n",
      "\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:355: RuntimeWarning:\n",
      "\n",
      "invalid value encountered in matmul\n",
      "\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:577: RuntimeWarning:\n",
      "\n",
      "divide by zero encountered in matmul\n",
      "\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:577: RuntimeWarning:\n",
      "\n",
      "overflow encountered in matmul\n",
      "\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:577: RuntimeWarning:\n",
      "\n",
      "invalid value encountered in matmul\n",
      "\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:590: RuntimeWarning:\n",
      "\n",
      "divide by zero encountered in matmul\n",
      "\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:590: RuntimeWarning:\n",
      "\n",
      "overflow encountered in matmul\n",
      "\n",
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/sklearn/utils/extmath.py:590: RuntimeWarning:\n",
      "\n",
      "invalid value encountered in matmul\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Reduce embeddings to 3D with t-SNE\n",
    "tsne = TSNE(n_components=3, random_state=42, perplexity=30)\n",
    "embeddings_3d = tsne.fit_transform(np.vstack(df[\"embedding\"].values))\n",
    "\n",
    "# Add t-SNE components to dataframe\n",
    "df[\"tsne1\"] = embeddings_3d[:, 0]\n",
    "df[\"tsne2\"] = embeddings_3d[:, 1]\n",
    "df[\"tsne3\"] = embeddings_3d[:, 2]\n",
    "\n",
    "# Interactive 3D scatter plot\n",
    "fig = px.scatter_3d(\n",
    "    df,\n",
    "    x=\"tsne1\",\n",
    "    y=\"tsne2\",\n",
    "    z=\"tsne3\",\n",
    "    color=\"age_recoded_6_categories\",\n",
    "    hover_data=[\"user_id\", \"gender\", \"country_code_iso_3166\"],\n",
    "    title=\"User Embeddings (t-SNE 3D)\"\n",
    ")\n",
    "fig.write_html(\"./output/user_embeddings_tsne_3d.html\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0043ad87-c0ee-4fe7-be3f-fe58ad9925bc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[{'user_id': '9561_FR_55-64_Man',\n",
       "  'user_text': \"risks_cntry_most_exposed_to_firstly: Cybersecurity threats (e.g. cyberattacks, cybercrimes, etc.) | risks_pers_most_exposed_to_firstly: Nuclear accidents | risks_pers_most_exposed_to_number_of_mentioned_risks: 2 mentions | pot_info_sources_to_learn_about_disaster_risks_firstly: National media | statements_disaster_risks_readseenheard_info_in_last_12m: Tend to agree | statements_disaster_risks_feel_well_informed: Tend to agree | statements_disaster_risks_trust_information_by_pub_auth_on_risks_where_you_live: Tend to disagree | statements_disaster_risks_easy_to_find_information_by_pub_auth_on_risks_where_you_live: Tend to agree | statements_disaster_risks_know_where_to_find_info_when_travelling_to_oth_eu_cntry: Don't know (SPONTANEOUS) | disaster_measures_in_hh_emergency_supply_drinksfood: Not mentioned | disaster_measures_in_hh_emergency_supply_water_cookinghygiene: Not mentioned | disaster_measures_in_hh_flashlightcandles: Not mentioned | disaster_measures_in_hh_batterypowered_radio: Not mentioned | disaster_measures_in_hh_emergency_pharmacy: Not mentioned | disaster_measures_in_hh_copies_imp_documentsstored_safely: Not mentioned | disaster_measures_in_hh_emergency_grabbag: Not mentioned | disaster_measures_in_hh_signed_up_for_alerts: Not mentioned | disaster_measures_in_hh_participated_in_trainingexercise: Not mentioned | disaster_measures_in_hh_informed_about_official_response_plan: Not mentioned | disaster_measures_in_hh_agreed_with_friendsfamily_to_contact: Not mentioned | disaster_measures_in_hh_discussed_common_prot_measures_in_neighbourhood: Not mentioned | disaster_measures_in_hh_invested_in_prot_measures_in_home: Not mentioned | how_many_days_meet_water_needs_if_water_services_disrupted: 4 - 7 days | how_many_days_power_essent_appliances_if_elec_interrupted: 2 - 3 days | how_many_days_cook_mealsheat_if_gas_disrupted: 1 day or less | how_many_days_provide_food_if_transportation_disrupted: 2 - 3 days | how_many_days_continued_treatment_if_medication_supply_disrupted: 2 - 3 days | personal_disaster_preparedness_better_able_to_cope_by_prep: Tend to agree | personal_disaster_preparedness_feel_well_prepared: Don't know (SPONTANEOUS) | personal_disaster_preparedness_no_timefin_resources_to_prep: Tend to agree | personal_disaster_preparedness_easy_to_find_info_on_how_to_prep: Don't know (SPONTANEOUS) | personal_disaster_preparedness_need_more_info_to_prep: Totally agree | personal_disaster_preparedness_know_how_emerg_services_will_alert: Tend to disagree | personal_disaster_preparedness_know_what_to_do_in_event_of_disaster: Don't know (SPONTANEOUS) | personal_disaster_preparedness_employerschool_encourages_trainingprep: It depends on the type of disaster (SPONTANEOUS) | personal_disaster_preparedness_emerg_services_encourage_trainingprep: Don't know (SPONTANEOUS) | relying_on_in_first_days_of_disaster_familyfriends: Somewhat | relying_on_in_first_days_of_disaster_people_in_neighbourhood: Somewhat | relying_on_in_first_days_of_disaster_assocsnonprofit_orgs: Somewhat | relying_on_in_first_days_of_disaster_emerg_services: Somewhat | relying_on_in_first_days_of_disaster_local_authgvmt_services: Somewhat | relying_on_in_first_days_of_disaster_workemployerschooledu_institution: Somewhat | relying_on_in_first_days_of_disaster_private_sector_entities: Somewhat | trust_in_emerg_services_to_handle_disastersemerg_situations_properly: Tend to trust | engaging_in_voluntary_work_for_emerg_responder_orgs: No, you have never engaged in voluntary work and do not plan to do so | occupation_of_respondent: Unemployed, temporarily not working | age_recoded_6_categories: 55-64 | size_of_community: Towns/suburbs | direction_things_are_going_life_personally: Things are going in the wrong direction | political_discussion_local_matters: Occasionally | political_discussion_national_matters: Occasionally | internet_use_total: Everyday/almost everyday (at least once 1 in d62_1 to d62_4) | gender: Man | age_education: 16.0 | standard_of_living_last_5yrs_in_light_of_crises: Your standard of living has decreased | personal_living_conditions_in_one_years_time: Worse | standard_of_living_next_5yrs: Your standard of living will decrease | region_austria: nan | region_belgium: nan | region_bulgaria: nan | region_croatia: nan | region_cyprus: nan | region_czechia: nan | region_denmark: nan | region_germany: nan | region_estonia: nan | region_finland: nan | region_france: FRK2 Rhone-Alpes | region_greece: nan | region_hungary: nan | region_ireland: nan | region_italy: nan | region_latvia: nan | region_lithuania: nan | region_luxembourg: nan | region_malta: nan | region_netherlands: nan | region_poland: nan | region_portugal: nan | region_romania: nan | region_slovenia: nan | region_slovakia: nan | region_spain: nan | region_sweden: nan | education_level_preprimary_education_incl_no_education: Not mentioned | education_level_primary_education: Not mentioned | education_level_lower_secondary_education: Lower secondary education | education_level_upper_secondary_education: Not mentioned | education_level_postsecondary_non_tertiary_incl_prevocationalvocational: Not mentioned | education_level_shortcycle_tertiary: Not mentioned | education_level_bachelor_or_equivalent: Not mentioned | education_level_master_or_equivalent: Not mentioned | education_level_doctoral_or_equivalent: Not mentioned\",\n",
       "  'similarity': 0.8321311374074787},\n",
       " {'user_id': '5096_DE-W_65-74_Woman',\n",
       "  'user_text': 'risks_cntry_most_exposed_to_firstly: Terrorist attacks | risks_pers_most_exposed_to_firstly: Extreme weather events (violent storms, droughts, heatwaves, cold waves, etc.) | risks_pers_most_exposed_to_number_of_mentioned_risks: +5 mentions | pot_info_sources_to_learn_about_disaster_risks_firstly: Emergency management services (e.g. police, firefighters, civil protection) | statements_disaster_risks_readseenheard_info_in_last_12m: Tend to agree | statements_disaster_risks_feel_well_informed: Tend to agree | statements_disaster_risks_trust_information_by_pub_auth_on_risks_where_you_live: Totally agree | statements_disaster_risks_easy_to_find_information_by_pub_auth_on_risks_where_you_live: Tend to disagree | statements_disaster_risks_know_where_to_find_info_when_travelling_to_oth_eu_cntry: Totally disagree | disaster_measures_in_hh_emergency_supply_drinksfood: Keep an emergency supply stock/pack of drinks, food | disaster_measures_in_hh_emergency_supply_water_cookinghygiene: Keep an emergency supply of water for cooking and hygiene | disaster_measures_in_hh_flashlightcandles: Have flashlight or candles accessible | disaster_measures_in_hh_batterypowered_radio: Not mentioned | disaster_measures_in_hh_emergency_pharmacy: Not mentioned | disaster_measures_in_hh_copies_imp_documentsstored_safely: Have made sure you have copies of your most important documents or have stored them safely | disaster_measures_in_hh_emergency_grabbag: Not mentioned | disaster_measures_in_hh_signed_up_for_alerts: Not mentioned | disaster_measures_in_hh_participated_in_trainingexercise: Not mentioned | disaster_measures_in_hh_informed_about_official_response_plan: Not mentioned | disaster_measures_in_hh_agreed_with_friendsfamily_to_contact: Agreed with family, friends on how to contact each other in case of an emergency | disaster_measures_in_hh_discussed_common_prot_measures_in_neighbourhood: Not mentioned | disaster_measures_in_hh_invested_in_prot_measures_in_home: Have invested in protective measures in your home (e.g. flood-proofed the electricity installation, cleared (...) | how_many_days_meet_water_needs_if_water_services_disrupted: 4 - 7 days | how_many_days_power_essent_appliances_if_elec_interrupted: 1 day or less | how_many_days_cook_mealsheat_if_gas_disrupted: 2 - 3 days | how_many_days_provide_food_if_transportation_disrupted: 4 - 7 days | how_many_days_continued_treatment_if_medication_supply_disrupted: 4 - 7 days | personal_disaster_preparedness_better_able_to_cope_by_prep: Tend to agree | personal_disaster_preparedness_feel_well_prepared: Tend to agree | personal_disaster_preparedness_no_timefin_resources_to_prep: Totally disagree | personal_disaster_preparedness_easy_to_find_info_on_how_to_prep: It depends on the type of disaster (SPONTANEOUS) | personal_disaster_preparedness_need_more_info_to_prep: Tend to agree | personal_disaster_preparedness_know_how_emerg_services_will_alert: Totally agree | personal_disaster_preparedness_know_what_to_do_in_event_of_disaster: Totally agree | personal_disaster_preparedness_employerschool_encourages_trainingprep: Totally disagree | personal_disaster_preparedness_emerg_services_encourage_trainingprep: Tend to agree | relying_on_in_first_days_of_disaster_familyfriends: A great deal | relying_on_in_first_days_of_disaster_people_in_neighbourhood: Somewhat | relying_on_in_first_days_of_disaster_assocsnonprofit_orgs: Somewhat | relying_on_in_first_days_of_disaster_emerg_services: Somewhat | relying_on_in_first_days_of_disaster_local_authgvmt_services: Somewhat | relying_on_in_first_days_of_disaster_workemployerschooledu_institution: Not at all | relying_on_in_first_days_of_disaster_private_sector_entities: Not a lot | trust_in_emerg_services_to_handle_disastersemerg_situations_properly: Totally trust | engaging_in_voluntary_work_for_emerg_responder_orgs: No, you have never engaged in voluntary work and do not plan to do so | occupation_of_respondent: Retired, unable to work | age_recoded_6_categories: 65-74 | size_of_community: Rural | direction_things_are_going_life_personally: Things are going in the right direction | political_discussion_local_matters: Frequently | political_discussion_national_matters: Frequently | internet_use_total: Often/sometimes (at least once 2-5 in d62_1 to d62_4) | gender: Woman | age_education: 17.0 | standard_of_living_last_5yrs_in_light_of_crises: Your standard of living has not changed | personal_living_conditions_in_one_years_time: Better | standard_of_living_next_5yrs: Your standard of living will not change | region_austria: nan | region_belgium: nan | region_bulgaria: nan | region_croatia: nan | region_cyprus: nan | region_czechia: nan | region_denmark: nan | region_germany: DE9 - Niedersachsen | region_estonia: nan | region_finland: nan | region_france: nan | region_greece: nan | region_hungary: nan | region_ireland: nan | region_italy: nan | region_latvia: nan | region_lithuania: nan | region_luxembourg: nan | region_malta: nan | region_netherlands: nan | region_poland: nan | region_portugal: nan | region_romania: nan | region_slovenia: nan | region_slovakia: nan | region_spain: nan | region_sweden: nan | education_level_preprimary_education_incl_no_education: Not mentioned | education_level_primary_education: Not mentioned | education_level_lower_secondary_education: Lower secondary education | education_level_upper_secondary_education: Not mentioned | education_level_postsecondary_non_tertiary_incl_prevocationalvocational: Not mentioned | education_level_shortcycle_tertiary: Not mentioned | education_level_bachelor_or_equivalent: Not mentioned | education_level_master_or_equivalent: Not mentioned | education_level_doctoral_or_equivalent: Not mentioned',\n",
       "  'similarity': 0.8967985948973097}]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# example comparison based on t-sne projection\n",
    "index_1 = 9561\n",
    "index_2 = 5096\n",
    "df.loc[[index_1, index_2]][[\"user_id\", \"user_text\", \"similarity\"]].to_dict(orient=\"records\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "967ab8ce-1769-4b84-a1ed-9d2cc6e12faa",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "disaster_measures_in_hh_emergency_supply_drinksfood: Not mentioned | Keep an emergency supply stock/pack of drinks, food\n",
      "disaster_measures_in_hh_emergency_supply_water_cookinghygiene: Not mentioned | Keep an emergency supply of water for cooking and hygiene\n",
      "disaster_measures_in_hh_flashlightcandles: Not mentioned | Have flashlight or candles accessible\n",
      "disaster_measures_in_hh_copies_imp_documentsstored_safely: Not mentioned | Have made sure you have copies of your most important documents or have stored them safely\n",
      "disaster_measures_in_hh_agreed_with_friendsfamily_to_contact: Not mentioned | Agreed with family, friends on how to contact each other in case of an emergency\n",
      "disaster_measures_in_hh_invested_in_prot_measures_in_home: Not mentioned | Have invested in protective measures in your home (e.g. flood-proofed the electricity installation, cleared (...)\n"
     ]
    }
   ],
   "source": [
    "row1 = df.loc[index_1].drop(labels=[\"user_id\", \"embedding\", \"user_text\", \"similarity\", \"pca1\", \"pca2\", \"pca3\", \"tsne1\", \"tsne2\", \"tsne3\"])\n",
    "row2 = df.loc[index_2].drop(labels=[\"user_id\", \"embedding\", \"user_text\", \"similarity\", \"pca1\", \"pca2\", \"pca3\", \"tsne1\", \"tsne2\", \"tsne3\"])\n",
    "# column_list = df.columns.difference([\"user_id\", \"embedding\", \"user_text\", \"similarity\", \"pca1\", \"pca2\", \"pca3\", \"tsne1\", \"tsne2\", \"tsne3\"])\n",
    "column_list = disaster_measures_in_hh_columns + how_many_days_meet_columns\n",
    "\n",
    "diffs = {}\n",
    "for col in column_list:\n",
    "    val1 = row1[col]\n",
    "    val2 = row2[col]\n",
    "    if pd.isnull(val1) and pd.isnull(val2):\n",
    "        continue\n",
    "    if val1 != val2:\n",
    "        diffs[col] = (val1, val2)\n",
    "\n",
    "# Print or display the differing columns and their values\n",
    "for col, (v1, v2) in diffs.items():\n",
    "    print(f\"{col}: {v1} | {v2}\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "964a3ca5-0ba3-4205-9553-6f2a16007b47",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/Users/jamestwose/Coding/matti_jms_collabs/.venv/lib/python3.12/site-packages/umap/umap_.py:1952: UserWarning:\n",
      "\n",
      "n_jobs value 1 overridden to 1 by setting random_state. Use no seed for parallelism.\n",
      "\n"
     ]
    }
   ],
   "source": [
    "# Reduce embeddings to 3D with UMAP\n",
    "umap_3d = umap.UMAP(n_components=3, random_state=42, metric=\"cosine\").fit_transform(np.vstack(df[\"embedding\"].values))\n",
    "\n",
    "# Add UMAP components to dataframe\n",
    "df[\"umap1\"] = umap_3d[:, 0]\n",
    "df[\"umap2\"] = umap_3d[:, 1]\n",
    "df[\"umap3\"] = umap_3d[:, 2]\n",
    "\n",
    "# Interactive 3D scatter plot\n",
    "fig = px.scatter_3d(\n",
    "    df,\n",
    "    x=\"umap1\",\n",
    "    y=\"umap2\",\n",
    "    z=\"umap3\",\n",
    "    color=\"country_code_iso_3166\",\n",
    "    hover_data=[\"user_id\", \"gender\", \"country_code_iso_3166\"],\n",
    "    title=\"User Embeddings (UMAP 3D)\"\n",
    ")\n",
    "fig.write_html(\"./output/user_embeddings_umap_3d.html\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "415c159f-5dd5-473b-8c7d-1b3d76429145",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[{'user_id': '21396_PT_45-54_Man',\n",
       "  'user_text': 'risks_cntry_most_exposed_to_firstly: Extreme weather events (violent storms, droughts, heatwaves, cold waves, etc.) | risks_pers_most_exposed_to_firstly: Extreme weather events (violent storms, droughts, heatwaves, cold waves, etc.) | risks_pers_most_exposed_to_number_of_mentioned_risks: 3 mentions | pot_info_sources_to_learn_about_disaster_risks_firstly: Emergency management services (e.g. police, firefighters, civil protection) | statements_disaster_risks_readseenheard_info_in_last_12m: Tend to disagree | statements_disaster_risks_feel_well_informed: Tend to disagree | statements_disaster_risks_trust_information_by_pub_auth_on_risks_where_you_live: Tend to agree | statements_disaster_risks_easy_to_find_information_by_pub_auth_on_risks_where_you_live: Tend to disagree | statements_disaster_risks_know_where_to_find_info_when_travelling_to_oth_eu_cntry: Tend to disagree | disaster_measures_in_hh_emergency_supply_drinksfood: Not mentioned | disaster_measures_in_hh_emergency_supply_water_cookinghygiene: Not mentioned | disaster_measures_in_hh_flashlightcandles: Have flashlight or candles accessible | disaster_measures_in_hh_batterypowered_radio: Not mentioned | disaster_measures_in_hh_emergency_pharmacy: Keep a home pharmacy for emergencies | disaster_measures_in_hh_copies_imp_documentsstored_safely: Not mentioned | disaster_measures_in_hh_emergency_grabbag: Not mentioned | disaster_measures_in_hh_signed_up_for_alerts: Not mentioned | disaster_measures_in_hh_participated_in_trainingexercise: Not mentioned | disaster_measures_in_hh_informed_about_official_response_plan: Not mentioned | disaster_measures_in_hh_agreed_with_friendsfamily_to_contact: Not mentioned | disaster_measures_in_hh_discussed_common_prot_measures_in_neighbourhood: Not mentioned | disaster_measures_in_hh_invested_in_prot_measures_in_home: Not mentioned | how_many_days_meet_water_needs_if_water_services_disrupted: 1 day or less | how_many_days_power_essent_appliances_if_elec_interrupted: 2 - 3 days | how_many_days_cook_mealsheat_if_gas_disrupted: 1 day or less | how_many_days_provide_food_if_transportation_disrupted: 4 - 7 days | how_many_days_continued_treatment_if_medication_supply_disrupted: More than 7 days | personal_disaster_preparedness_better_able_to_cope_by_prep: Tend to agree | personal_disaster_preparedness_feel_well_prepared: Tend to disagree | personal_disaster_preparedness_no_timefin_resources_to_prep: Tend to agree | personal_disaster_preparedness_easy_to_find_info_on_how_to_prep: Tend to disagree | personal_disaster_preparedness_need_more_info_to_prep: Tend to agree | personal_disaster_preparedness_know_how_emerg_services_will_alert: Tend to disagree | personal_disaster_preparedness_know_what_to_do_in_event_of_disaster: Tend to disagree | personal_disaster_preparedness_employerschool_encourages_trainingprep: Tend to disagree | personal_disaster_preparedness_emerg_services_encourage_trainingprep: Tend to disagree | relying_on_in_first_days_of_disaster_familyfriends: A great deal | relying_on_in_first_days_of_disaster_people_in_neighbourhood: Not a lot | relying_on_in_first_days_of_disaster_assocsnonprofit_orgs: Somewhat | relying_on_in_first_days_of_disaster_emerg_services: A great deal | relying_on_in_first_days_of_disaster_local_authgvmt_services: A great deal | relying_on_in_first_days_of_disaster_workemployerschooledu_institution: Not at all | relying_on_in_first_days_of_disaster_private_sector_entities: Not a lot | trust_in_emerg_services_to_handle_disastersemerg_situations_properly: Tend to trust | engaging_in_voluntary_work_for_emerg_responder_orgs: No, you have never engaged in voluntary work and do not plan to do so | occupation_of_respondent: Employed position, travelling | age_recoded_6_categories: 45-54 | size_of_community: Rural | direction_things_are_going_life_personally: Neither the one nor the other (SPONTANEOUS) | political_discussion_local_matters: Never | political_discussion_national_matters: Never | internet_use_total: Everyday/almost everyday (at least once 1 in d62_1 to d62_4) | gender: Man | age_education: 15.0 | standard_of_living_last_5yrs_in_light_of_crises: Your standard of living has decreased | personal_living_conditions_in_one_years_time: Worse | standard_of_living_next_5yrs: Your standard of living will decrease | region_austria: nan | region_belgium: nan | region_bulgaria: nan | region_croatia: nan | region_cyprus: nan | region_czechia: nan | region_denmark: nan | region_germany: nan | region_estonia: nan | region_finland: nan | region_france: nan | region_greece: nan | region_hungary: nan | region_ireland: nan | region_italy: nan | region_latvia: nan | region_lithuania: nan | region_luxembourg: nan | region_malta: nan | region_netherlands: nan | region_poland: nan | region_portugal: PT16 - Centro (PT) | region_romania: nan | region_slovenia: nan | region_slovakia: nan | region_spain: nan | region_sweden: nan | education_level_preprimary_education_incl_no_education: Not mentioned | education_level_primary_education: Not mentioned | education_level_lower_secondary_education: Lower secondary education | education_level_upper_secondary_education: Not mentioned | education_level_postsecondary_non_tertiary_incl_prevocationalvocational: Not mentioned | education_level_shortcycle_tertiary: Not mentioned | education_level_bachelor_or_equivalent: Not mentioned | education_level_master_or_equivalent: Not mentioned | education_level_doctoral_or_equivalent: Not mentioned',\n",
       "  'similarity': 0.8832531753098408},\n",
       " {'user_id': '10444_FR_65-74_Woman',\n",
       "  'user_text': \"risks_cntry_most_exposed_to_firstly: Extreme weather events (violent storms, droughts, heatwaves, cold waves, etc.) | risks_pers_most_exposed_to_firstly: Extreme weather events (violent storms, droughts, heatwaves, cold waves, etc.) | risks_pers_most_exposed_to_number_of_mentioned_risks: 4 mentions | pot_info_sources_to_learn_about_disaster_risks_firstly: Local or national authorities or agencies | statements_disaster_risks_readseenheard_info_in_last_12m: Tend to agree | statements_disaster_risks_feel_well_informed: Totally disagree | statements_disaster_risks_trust_information_by_pub_auth_on_risks_where_you_live: Tend to agree | statements_disaster_risks_easy_to_find_information_by_pub_auth_on_risks_where_you_live: Totally disagree | statements_disaster_risks_know_where_to_find_info_when_travelling_to_oth_eu_cntry: Totally disagree | disaster_measures_in_hh_emergency_supply_drinksfood: Not mentioned | disaster_measures_in_hh_emergency_supply_water_cookinghygiene: Not mentioned | disaster_measures_in_hh_flashlightcandles: Not mentioned | disaster_measures_in_hh_batterypowered_radio: Not mentioned | disaster_measures_in_hh_emergency_pharmacy: Not mentioned | disaster_measures_in_hh_copies_imp_documentsstored_safely: Not mentioned | disaster_measures_in_hh_emergency_grabbag: Not mentioned | disaster_measures_in_hh_signed_up_for_alerts: Not mentioned | disaster_measures_in_hh_participated_in_trainingexercise: Not mentioned | disaster_measures_in_hh_informed_about_official_response_plan: Not mentioned | disaster_measures_in_hh_agreed_with_friendsfamily_to_contact: Not mentioned | disaster_measures_in_hh_discussed_common_prot_measures_in_neighbourhood: Not mentioned | disaster_measures_in_hh_invested_in_prot_measures_in_home: Not mentioned | how_many_days_meet_water_needs_if_water_services_disrupted: 2 - 3 days | how_many_days_power_essent_appliances_if_elec_interrupted: 1 day or less | how_many_days_cook_mealsheat_if_gas_disrupted: 1 day or less | how_many_days_provide_food_if_transportation_disrupted: 2 - 3 days | how_many_days_continued_treatment_if_medication_supply_disrupted: Don't know (SPONTANEOUS) | personal_disaster_preparedness_better_able_to_cope_by_prep: Tend to agree | personal_disaster_preparedness_feel_well_prepared: Totally disagree | personal_disaster_preparedness_no_timefin_resources_to_prep: Don't know (SPONTANEOUS) | personal_disaster_preparedness_easy_to_find_info_on_how_to_prep: Totally disagree | personal_disaster_preparedness_need_more_info_to_prep: Tend to agree | personal_disaster_preparedness_know_how_emerg_services_will_alert: Totally disagree | personal_disaster_preparedness_know_what_to_do_in_event_of_disaster: Totally disagree | personal_disaster_preparedness_employerschool_encourages_trainingprep: Don't know (SPONTANEOUS) | personal_disaster_preparedness_emerg_services_encourage_trainingprep: Don't know (SPONTANEOUS) | relying_on_in_first_days_of_disaster_familyfriends: A great deal | relying_on_in_first_days_of_disaster_people_in_neighbourhood: A great deal | relying_on_in_first_days_of_disaster_assocsnonprofit_orgs: A great deal | relying_on_in_first_days_of_disaster_emerg_services: A great deal | relying_on_in_first_days_of_disaster_local_authgvmt_services: A great deal | relying_on_in_first_days_of_disaster_workemployerschooledu_institution: Don't know (SPONTANEOUS) | relying_on_in_first_days_of_disaster_private_sector_entities: Don't know (SPONTANEOUS) | trust_in_emerg_services_to_handle_disastersemerg_situations_properly: Tend to trust | engaging_in_voluntary_work_for_emerg_responder_orgs: No, but you have engaged in voluntary work in the past | occupation_of_respondent: Retired, unable to work | age_recoded_6_categories: 65-74 | size_of_community: Cities/urban | direction_things_are_going_life_personally: DK (SPONT.) | political_discussion_local_matters: Occasionally | political_discussion_national_matters: Occasionally | internet_use_total: Often/sometimes (at least once 2-5 in d62_1 to d62_4) | gender: Woman | age_education: 17.0 | standard_of_living_last_5yrs_in_light_of_crises: Your standard of living has decreased | personal_living_conditions_in_one_years_time: Worse | standard_of_living_next_5yrs: Your standard of living will decrease | region_austria: nan | region_belgium: nan | region_bulgaria: nan | region_croatia: nan | region_cyprus: nan | region_czechia: nan | region_denmark: nan | region_germany: nan | region_estonia: nan | region_finland: nan | region_france: FR10 Ile-de-France | region_greece: nan | region_hungary: nan | region_ireland: nan | region_italy: nan | region_latvia: nan | region_lithuania: nan | region_luxembourg: nan | region_malta: nan | region_netherlands: nan | region_poland: nan | region_portugal: nan | region_romania: nan | region_slovenia: nan | region_slovakia: nan | region_spain: nan | region_sweden: nan | education_level_preprimary_education_incl_no_education: Not mentioned | education_level_primary_education: Not mentioned | education_level_lower_secondary_education: Lower secondary education | education_level_upper_secondary_education: Not mentioned | education_level_postsecondary_non_tertiary_incl_prevocationalvocational: Not mentioned | education_level_shortcycle_tertiary: Not mentioned | education_level_bachelor_or_equivalent: Not mentioned | education_level_master_or_equivalent: Not mentioned | education_level_doctoral_or_equivalent: Not mentioned\",\n",
       "  'similarity': 0.8832531753098408}]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# example comparison based on UMAP projection\n",
    "index_1 = 21396\n",
    "index_2 = 10444\n",
    "df.loc[[index_1, index_2]][[\"user_id\", \"user_text\", \"similarity\"]].to_dict(orient=\"records\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "24aa9cca-7c83-4cf1-a256-562c1f9bdc81",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "disaster_measures_in_hh_flashlightcandles: Have flashlight or candles accessible | Not mentioned\n",
      "disaster_measures_in_hh_emergency_pharmacy: Keep a home pharmacy for emergencies | Not mentioned\n",
      "how_many_days_meet_water_needs_if_water_services_disrupted: 1 day or less | 2 - 3 days\n"
     ]
    }
   ],
   "source": [
    "row1 = df.loc[index_1].drop(labels=[\"user_id\", \"embedding\", \"user_text\", \"similarity\", \"pca1\", \"pca2\", \"pca3\", \"tsne1\", \"tsne2\", \"tsne3\", \"umap1\", \"umap2\", \"umap3\"])\n",
    "row2 = df.loc[index_2].drop(labels=[\"user_id\", \"embedding\", \"user_text\", \"similarity\", \"pca1\", \"pca2\", \"pca3\", \"tsne1\", \"tsne2\", \"tsne3\", \"umap1\", \"umap2\", \"umap3\"])\n",
    "# column_list = df.columns.difference([\"user_id\", \"embedding\", \"user_text\", \"similarity\", \"pca1\", \"pca2\", \"pca3\", \"tsne1\", \"tsne2\", \"tsne3\", \"umap1\", \"umap2\", \"umap3\"])\n",
    "column_list = disaster_measures_in_hh_columns + how_many_days_meet_columns\n",
    "diffs = {}\n",
    "for col in column_list:\n",
    "    val1 = row1[col]\n",
    "    val2 = row2[col]\n",
    "    if pd.isnull(val1) and pd.isnull(val2):\n",
    "        continue\n",
    "    if val1 != val2:\n",
    "        diffs[col] = (val1, val2)\n",
    "\n",
    "# Print or display the differing columns and their values\n",
    "for col, (v1, v2) in diffs.items():\n",
    "    print(f\"{col}: {v1} | {v2}\")"
   ]
  }
 ],
 "metadata": {
  "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
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}