tutorial/homophily.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Analysing homophily based on two different null models\n",
    "\n",
    "**Authors**: Andreas Kruff, Johann Schaible, Marcos Oliveira\n",
    "\n",
    "**Version**: 20.04.2020\n",
    "\n",
    "**Description**: To analyze the connectivity of a group of persons (at a conference) for example, this package offers you the possibility to apply two different null models."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The data sets that are used in this tutorial are part of the following paper:\n",
    "\n",
    "**Génois, Mathieu & Zens, Maria & Lechner, Clemens & Rammstedt, Beatrice & Strohmaier, Markus. (2019). Building connections: How scientists meet each other during a conference.**\n",
    "\n",
    "The data sets are available here: \n",
    "\n",
    "https://zenodo.org/record/2531537#.X0OObcgzaUl"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For more information about the methods that are explained in this tutorial you can check out the online documentation of this toolbox here:\n",
    "\n",
    "https://gesiscss.github.io/face2face/"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Table of Contents\n",
    "#### [Apply Configuration model on a network](#configuration_model)\n",
    "#### [Plotting histogram for the null model](#histogram)\n",
    "#### [Apply null model with randomized identities](#null_model)\n",
    "#### [Plotting a heatmap from contact matrix](#plot_heatmap)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "import face2face as f2f\n",
    "\n",
    "df = f2f.Data(\"WS16\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The homophily describes the tendency that individuals prefer to interact with other individuals that have similar attributes in terms of the socio-demographic informations for example. For the further analysis of this phenomenon the toolbox provides two null models that randomize the given networks in various ways to test different effects. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Apply Configuration Model on a network\n",
    "<a name=\"configuration_model\"></a>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The configuration model also known as the Sneppen-Maslov rewiring is a null model. A null model is a model which is based on a certain hypothesis about the interaction behaviour for example, which then needs to be tested by statistical methods. \n",
    "\n",
    "Configuration model:\n",
    "- randomized connections/links for the participants/nodes\n",
    "- amount of contact from a person stays the same/ node degree stays the same\n",
    "- (amount of contacts between two individuals/link weights get redistributed) \n",
    "\n",
    "In this case the null model all connections between the participants become randomized while every participant still has the same amount of contacts. This null model allows us to compare original scenario with a scenario where the participants behave the same in terms of their connectivity, but with randomized contacts. So we can check if two different groups are more or less connected which each other. If you want to learn more about the methods you can have a look at this paper:\n",
    "\n",
    "**Génois, Mathieu & Zens, Maria & Lechner, Clemens & Rammstedt, Beatrice & Strohmaier, Markus. (2019). Building connections: How scientists meet each other during a conference.**"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If you want to apply the Configuration model on the data set you can use the function \"configuration_model_label_z_score_mixing_matrix\" and you get the contact matrix with the z-scores.\n",
    "The function has a total of seven parameters. At first you need an object from type \"Data\". The more often you execute the function based on the \"runs\" parameter, the better the result should be in the end, because it is an heuristic approach. With the label parameter you can choose for which attribute you want to apply the null model. \n",
    "With the \"shuffle_label\" parameter you can extend the model by randomizing the peoples attributes. The \"force_simple_graph\"-parameter allows us to delete self edges which occur when we use the \"configuration_model\" function from networkx."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There are two additional (optional) input parameter \"seed_config_mat\" and \"seed_label\". You can use them if you need fixed results from this function. The input parameters have to be lists with an integer for every run you want to apply."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.0053674616717123,  0.5541368009847846,  0.9529846375336652,\n",
       "         0.083402634787837 ,  0.5251088044262069, -0.8244800545175114,\n",
       "        -0.9260490841194826],\n",
       "       [ 0.5541368009847846, -0.0933541252884789,  0.6392232710902084,\n",
       "         0.1955006696492974,  1.3876807443761165, -2.380567469175161 ,\n",
       "        -2.0567514302366185],\n",
       "       [ 0.9529846375336652,  0.6392232710902084,  1.2001088122261832,\n",
       "         1.3584915832692943,  1.0253397170100398, -1.5385526648824752,\n",
       "        -3.2085954565290282],\n",
       "       [ 0.083402634787837 ,  0.1955006696492974,  1.3584915832692943,\n",
       "         0.5675247260204097,  2.004494751792777 , -0.8371259228789633,\n",
       "        -1.1243699427926301],\n",
       "       [ 0.5251088044262069,  1.3876807443761165,  1.0253397170100398,\n",
       "         2.004494751792777 ,  0.4123744371798425,  0.0442245719566842,\n",
       "        -0.2746397619595122],\n",
       "       [-0.8244800545175114, -2.380567469175161 , -1.5385526648824752,\n",
       "        -0.8371259228789633,  0.0442245719566842, -1.237082802890754 ,\n",
       "        -3.444537920188045 ],\n",
       "       [-0.9260490841194826, -2.0567514302366185, -3.2085954565290282,\n",
       "        -1.1243699427926301, -0.2746397619595122, -3.444537920188045 ,\n",
       "        -0.7437081695398294]])"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "contact_matrix, matrices , data_mixing_matrix, abs_matrices = f2f.configuration_model_label_z_score_mixing_matrix(df, runs=10, label=\"Education\", shuffle_label=True, force_simple_graph=True)\n",
    "\n",
    "contact_matrix"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The Configuration Model allows us to analyze if two different groups are significantly more or less connected with each other then with other groups, because we keep the same individual behaviour in connectivity, but we randomize the connections between the people. The most relevant output for you as a user is the contact_matrix. This array contains the z-score combinations based on the label attribute you have choosen as the input parameter. The z-score describes positive and negative deviations as the amount of standard deviations $\\sigma$ from the mean. For the connectivity we compare the link density with regard to the attribute label. What we get is a symmetric matrix with the same x and y axis order so on the diagonal we compare the connectivity within the same group while the other values describe the connectivity between the people of two different groups. So a high positive value indicates a connectivity that is a lot higher then the average and a negative value indicates the opposite.\n",
    "\n",
    "The other output parameters are more relevant for the underlying functions behind the calculation and for the plotting, so if you just want to calculate and plot the homophily you can ignore them."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Plotting histogram for the null model\n",
    "<a name=\"histogram\"></a>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To visualize the contact matrix for a better comparison of the different expressions for the choosen label/attribute you can use the plot_null_model_subplots function. In the following figure you can see the overall z-score for the different combinations of expressions as a single darkgreen bar plot. The lightgreen bar plots show the amount of edges between the combinations. The input parameter are equal to the one from the null model functions. You just have to choose which null model you want to use by setting the model parameter. You can choose between \"configuration_model\" and \"random_identities\". If you choose \"configuration_model\" you can choose if you want to use the \"shuffle_label\" and the \"force_simple_graph\" parameter."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x288 with 8 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "f2f.plot_null_model_subplots(df, runs=100, label=\"Sex\",shuffle_label=True, force_simple_graph=True, model=\"configuration_model\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Apply null model with randomized identities\n",
    "<a name=\"null_model\"></a>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There is another null model you can apply on the dataset. The function is called \"shuffle_label_z_score_mixing_matrix\"\n",
    "and you get the contact matrix with the z-scores as output. Except for the \"shuffle_label\" and the \"force_simple_graph\" parameter the parameters are the same."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Null model 2:\n",
    "- randomizing the identities of the nodes/attributes of the participants\n",
    "- network structure stays the same\n",
    "\n",
    "If you want to learn more about the methods you can have a look at this paper:\n",
    "\n",
    "**Génois, Mathieu & Zens, Maria & Lechner, Clemens & Rammstedt, Beatrice & Strohmaier, Markus. (2019). Building connections: How scientists meet each other during a conference.**\n",
    "\n",
    "In this null model we want to compare if there are group pairs that significantly interact more or less when you keep the connectivity fixed, but you randomize the interactivity."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "contact_matrix, abs_matrices = f2f.shuffle_label_z_score_mixing_matrix(df, runs=10,label=\"Education\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ 0.5604485383178042,  0.7471608582404085,  1.4237922349132601,\n",
       "         0.8284961774187338,  1.0195857032052114, -0.5937006128972517,\n",
       "        -4.130921946615826 ],\n",
       "       [ 0.7471608582404085,  0.262311531293501 ,  0.3340704874951456,\n",
       "         0.5232218612309274,  1.3788910212595669, -1.8651130528588689,\n",
       "        -3.9898126085272265],\n",
       "       [ 1.4237922349132601,  0.3340704874951456,  0.7155646512351556,\n",
       "         1.6452498140154415,  0.8552913416991149, -2.264446845903306 ,\n",
       "        -4.53254266310579  ],\n",
       "       [ 0.8284961774187338,  0.5232218612309274,  1.6452498140154415,\n",
       "         0.8841118406457529,  1.7686858304188557, -1.344014173623075 ,\n",
       "        -3.7145674293345072],\n",
       "       [ 1.0195857032052114,  1.3788910212595669,  0.8552913416991149,\n",
       "         1.7686858304188557,  0.5419730101158714, -0.3341012065769349,\n",
       "        -2.828427124746193 ],\n",
       "       [-0.5937006128972517, -1.8651130528588689, -2.264446845903306 ,\n",
       "        -1.344014173623075 , -0.3341012065769349, -1.5179418517972914,\n",
       "        -7.276520484602838 ],\n",
       "       [-4.130921946615826 , -3.9898126085272265, -4.53254266310579  ,\n",
       "        -3.7145674293345072, -2.828427124746193 , -7.276520484602838 ,\n",
       "        -2.2360679774997907]])"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "contact_matrix"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "With this null model you can analyze if the correlations between topology and interactions are groupdependant or not. The output can be interpreted in the same way as the one from the configuration model."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Plotting a heatmap from contact matrix\n",
    "<a name=\"plot_heatmap\"></a>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "To plot the contact matrix you can use the \"plot_cm_heatmap\" function."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "f2f.plot_cm_heatmap(contact_matrix)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "contact_matrix, matrices , data_mixing_matrix, abs_matrices = f2f.configuration_model_label_z_score_mixing_matrix(df, runs=10, label=\"Sex\", shuffle_label=True, force_simple_graph=True)\n",
    "\n",
    "f2fplot_cm_heatmap(contact_matrix)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The heatmap visualization is based on the paper cited below:\n",
    "\n",
    "Génois, Mathieu & Zens, Maria & Lechner, Clemens & Rammstedt, Beatrice & Strohmaier, Markus. (2019). Building connections: How scientists meet each other during a conference. "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The Heatmap shows the z-score between every value-attribute combination. Red fields indicate positive deviations, blue fields indicate negative deviations. Z-Scores which are labeled with a white * are significant under p < 0.01 (taking into account a Bonferroni Correction). Huge outliers are pretty rare, which shows that the value-attributes mix rather well in this case."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}