assets/notebook/analysis_model_states.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "pgmYMbXxrRjn"
      },
      "outputs": [],
      "source": [
        "import os\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "from sklearn.preprocessing import StandardScaler\n",
        "import matplotlib.pyplot as plt\n",
        "from scipy.cluster.hierarchy import dendrogram, linkage, cophenet, fcluster\n",
        "from sklearn.cluster import AgglomerativeClustering\n",
        "from scipy.spatial.distance import pdist  # Import pdist\n",
        "from sklearn import metrics\n",
        "from scipy.spatial.distance import euclidean\n",
        "import json"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/drive')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nh3pxAujrfUT",
        "outputId": "8868e2f5-18c9-4c41-ba66-ade75620c406"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/drive\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "folder_path = '/content/drive/Shareddrives/DSCI554 project/Data collection/mental_health_data/data1_MHA_dashboard'\n",
        "filenames = os.listdir(folder_path)\n",
        "result_df = pd.read_csv(f'{folder_path}/depression_age_df.csv')\n",
        "result_df = result_df[['state','latitude','longitude']]\n",
        "for filename in filenames:\n",
        "    if '_age_' in filename:\n",
        "        df = pd.read_csv(os.path.join(folder_path, filename))\n",
        "        for col in df.columns:\n",
        "            if '2023-all_ages' in col and '100K' in col:\n",
        "                result_df = pd.merge(result_df,df[['state',f'{col}']],on=\"state\",how='outer')\n",
        "\n",
        "report_path = '/content/drive/Shareddrives/DSCI554 project/Data collection/mental_health_data/data2_MHA_report/2022_report'\n",
        "filenames = os.listdir(report_path)\n",
        "result_df2 = pd.read_csv(f'{report_path}/21_adults_with_substance_use_disorder_in_the_past_year.csv')\n",
        "result_df2 = result_df2[['state']]\n",
        "result_df2 = result_df2[result_df2['state'] != 'National']\n",
        "target_filenames = {'21_adults_with_substance_use_disorder_in_the_past_year.csv':'adult_disorder',\n",
        "                    '26_youth_with_substance_use_disorder_in_the_past_year.csv':'youth_disorder',\n",
        "                    '31_adults_with_cognitive_disability_who_could_not_see_a_doctor_due_to_costs.csv':'adult_cognitive_disability',\n",
        "                    '35_students_identified_with_emotional_disturbance_for_an_individualized_education_program.csv':'youth_cognitive_disability'}\n",
        "\n",
        "for filename in filenames:\n",
        "    if filename in target_filenames.keys():\n",
        "        df = pd.read_csv(os.path.join(report_path, filename))\n",
        "        df = df[['state','percent','number']]\n",
        "        df = df[df['state'] != 'National']\n",
        "        df.columns = ['state',f'{target_filenames[filename]}_percent',f'{target_filenames[filename]}_number']\n",
        "        df[f'{target_filenames[filename]}_percent'] = pd.to_numeric(df[f'{target_filenames[filename]}_percent'], errors='coerce')\n",
        "        df[f'{target_filenames[filename]}_number'] = pd.to_numeric(df[f'{target_filenames[filename]}_number'], errors='coerce')\n",
        "        df[f'{target_filenames[filename]}_total_number'] = (df[f'{target_filenames[filename]}_number']) / (0.01*df[f'{target_filenames[filename]}_percent'])\n",
        "        result_df2 = pd.merge(result_df2, df,on='state',how='outer')\n",
        "\n",
        "extracted_result_df2 = pd.DataFrame(result_df2['state'].copy())\n",
        "\n",
        "extracted_result_df2['cognitive_disability'] = 100*(\n",
        "    result_df2['youth_cognitive_disability_number'] + result_df2['adult_cognitive_disability_number']\n",
        ") / (\n",
        "    result_df2['youth_cognitive_disability_total_number'] + result_df2['adult_cognitive_disability_total_number']\n",
        ")\n",
        "\n",
        "extracted_result_df2['disorder'] = 100*(\n",
        "    result_df2['adult_disorder_number'] + result_df2['youth_disorder_number']\n",
        ") / (\n",
        "    result_df2['adult_disorder_total_number'] + result_df2['youth_cognitive_disability_number']\n",
        ")\n",
        "\n",
        "result_df = pd.merge(result_df,extracted_result_df2,on='state',how='outer')\n",
        "result_df.columns = ['state','latitude','longitude','depression','psychosis','ptsd','suicide','trauma','cognitive_disability','disorder']"
      ],
      "metadata": {
        "id": "WpK6S8TPsstd"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "radar chart - average performance on mental health features"
      ],
      "metadata": {
        "id": "kJ3fhXumZwJt"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "mental_health_features = result_df[['depression', 'psychosis', 'ptsd', 'suicide', 'trauma', 'cognitive_disability', 'disorder']]\n",
        "print(mental_health_features.describe())\n",
        "# Standardize the features\n",
        "scaler = StandardScaler()\n",
        "scaled_features = scaler.fit_transform(mental_health_features)\n",
        "\n",
        "# Perform hierarchical clustering\n",
        "linkage_method = 'ward'  # You can try other methods as well\n",
        "linkage_matrix = linkage(scaled_features, method=linkage_method)\n",
        "\n",
        "# Dendrogram\n",
        "plt.figure(figsize=(12, 6))\n",
        "plt.title('Hierarchical Clustering Dendrogram')\n",
        "dendrogram(linkage_matrix)\n",
        "plt.show()\n",
        "\n",
        "# Cophenetic correlation coefficient\n",
        "c, coph_dists = cophenet(linkage_matrix, pdist(scaled_features))\n",
        "print(f'Cophenetic Correlation Coefficient: {c:.3f}')\n",
        "\n",
        "# Calinski-Harabasz Index\n",
        "ch_scores = []\n",
        "for n_clusters in range(3, 11):\n",
        "    model = AgglomerativeClustering(n_clusters=n_clusters, linkage=linkage_method)\n",
        "    labels = model.fit_predict(scaled_features)\n",
        "    ch_score = metrics.calinski_harabasz_score(scaled_features, labels)\n",
        "    ch_scores.append(ch_score)\n",
        "\n",
        "# Plot Calinski-Harabasz Index\n",
        "plt.figure(figsize=(8, 4))\n",
        "plt.plot(range(3, 11), ch_scores, marker='o')\n",
        "plt.title('Calinski-Harabasz Index for Different Numbers of Clusters')\n",
        "plt.xlabel('Number of Clusters')\n",
        "plt.ylabel('Calinski-Harabasz Index')\n",
        "plt.show()\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "ctUWemMttU5o",
        "outputId": "5459797e-8f05-4b2b-988f-b3f9bf9434d2"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "       depression  psychosis       ptsd    suicide      trauma  \\\n",
            "count   51.000000  51.000000  51.000000  51.000000   51.000000   \n",
            "mean    30.668972  33.475846  13.422851  15.556723   70.347059   \n",
            "std      4.718872   5.198890   3.290769   3.136680   14.215912   \n",
            "min     21.950469  23.485920   7.154319   9.401156   42.000000   \n",
            "25%     27.897307  30.017426  10.791477  13.475686   60.050000   \n",
            "50%     30.516618  32.389270  13.494988  15.681878   72.300000   \n",
            "75%     32.888320  35.235183  15.657248  17.408386   80.250000   \n",
            "max     50.227045  56.368939  23.612171  25.659469  106.100000   \n",
            "\n",
            "       cognitive_disability   disorder  \n",
            "count             51.000000  51.000000  \n",
            "mean              25.506653   8.506061  \n",
            "std                4.133759   1.290740  \n",
            "min               17.761882   6.307285  \n",
            "25%               21.905363   7.632480  \n",
            "50%               25.260255   8.273193  \n",
            "75%               28.245895   9.170565  \n",
            "max               34.973245  12.633625  \n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1200x600 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Cophenetic Correlation Coefficient: 0.506\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 800x400 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "num_clusters = 4\n",
        "clusters = fcluster(linkage_matrix, t=num_clusters, criterion='maxclust')\n",
        "\n",
        "# Create a DataFrame with the original data and cluster assignments\n",
        "result_df_with_clusters = pd.concat([result_df, pd.DataFrame({'hcluster': clusters}, index=result_df.index)], axis=1)\n",
        "result_df_with_clusters\n",
        "import plotly.graph_objects as go\n",
        "result_df_with_clusters = result_df_with_clusters[['depression','psychosis','ptsd','suicide','cognitive_disability','disorder','hcluster']]\n",
        "cluster_avg = result_df_with_clusters.groupby('hcluster').mean()\n",
        "cluster_avg = pd.concat([cluster_avg.transpose(), pd.DataFrame(result_df_with_clusters.mean()[:-1])], axis=1).transpose()\n",
        "\n",
        "# Create a radar chart with multiple traces for each cluster\n",
        "fig = go.Figure()\n",
        "\n",
        "for cluster in cluster_avg.index:\n",
        "    fig.add_trace(go.Scatterpolar(\n",
        "        r=cluster_avg.loc[cluster].values,\n",
        "        theta=cluster_avg.columns,\n",
        "        fill='toself',\n",
        "        name=f'Cluster {cluster}'\n",
        "    ))\n",
        "\n",
        "fig.update_layout(\n",
        "    polar=dict(\n",
        "        radialaxis=dict(\n",
        "            visible=True,\n",
        "        ),\n",
        "    ),\n",
        "    showlegend=True,\n",
        "    title='Radar Chart for Clusters',\n",
        ")\n",
        "\n",
        "fig.show()"
      ],
      "metadata": {
        "id": "PBCjVzAEvsdf"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "d3_radar = dict()\n",
        "for cluster, row in cluster_avg.iterrows():\n",
        "    if cluster != 0:\n",
        "        d3_radar[cluster] = [{'axis': column, 'value': value} for column, value in row.items()]\n",
        "    else:\n",
        "        d3_radar['mean'] = [{'axis': column, 'value': value} for column, value in row.items()]\n",
        "\n",
        "json_data = json.dumps(d3_radar)\n",
        "with open('radar_data.json', 'w') as json_file:\n",
        "    json_file.write(json_data)"
      ],
      "metadata": {
        "id": "IFi8ea1QxGqQ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "network - prevalence, cluster distance"
      ],
      "metadata": {
        "id": "9xzHHnvOagcY"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "change_lat_lng = {42: [-72.4, 38.9],5: [-72.7678, 45.5584], 21: [-70.49, 43.66], 11: [-69.4757, 41.4966],44: [-72.7028, 40.9]}\n",
        "for state_id, coordinate in change_lat_lng.items():\n",
        "    result_df.loc[state_id,'latitude'] = coordinate[1]\n",
        "    result_df.loc[state_id,'longitude'] = coordinate[0]"
      ],
      "metadata": {
        "id": "HFb1V8YUbkr2"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "report_path = '/content/drive/Shareddrives/DSCI554 project/Data collection/mental_health_data/data2_MHA_report/2022_report'\n",
        "filenames = os.listdir(report_path)\n",
        "prevalence_df = pd.read_csv(f'{report_path}/20_adult_prevalence_of_mental_illness.csv')\n",
        "prevalence_df = prevalence_df[['state']]\n",
        "prevalence_df = prevalence_df[prevalence_df['state'] != 'National']\n",
        "target_filenames = {'20_adult_prevalence_of_mental_illness.csv':'adult_prevalence',\n",
        "                    '25_youth_prevalence_of_mental_illness.csv':'youth_prevalence'}\n",
        "for filename in filenames:\n",
        "    if filename in target_filenames.keys():\n",
        "        df = pd.read_csv(os.path.join(report_path, filename))\n",
        "        df = df[['state','percent','number']]\n",
        "        df = df[df['state'] != 'National']\n",
        "        df.columns = ['state',f'{target_filenames[filename]}_percent',f'{target_filenames[filename]}_number']\n",
        "        df[f'{target_filenames[filename]}_percent'] = pd.to_numeric(df[f'{target_filenames[filename]}_percent'], errors='coerce')\n",
        "        df[f'{target_filenames[filename]}_number'] = pd.to_numeric(df[f'{target_filenames[filename]}_number'], errors='coerce')\n",
        "        df[f'{target_filenames[filename]}_total_number'] = (df[f'{target_filenames[filename]}_number']) / (0.01*df[f'{target_filenames[filename]}_percent'])\n",
        "        prevalence_df = pd.merge(prevalence_df, df,on='state',how='outer')\n",
        "\n",
        "extracted_prevalence_df = pd.DataFrame(prevalence_df['state'].copy())\n",
        "\n",
        "extracted_prevalence_df['prevalence_percent'] = 100*(\n",
        "    prevalence_df['youth_prevalence_number'] + prevalence_df['adult_prevalence_number']\n",
        ") / (\n",
        "    prevalence_df['youth_prevalence_total_number'] + prevalence_df['adult_prevalence_total_number']\n",
        ")\n",
        "extracted_prevalence_df = pd.merge(result_df[['state','latitude','longitude']],extracted_prevalence_df,on='state', how ='outer')\n",
        "# extracted_prevalence_df"
      ],
      "metadata": {
        "id": "rIRIb9XPZqpE"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "cluster_means = []\n",
        "for cluster_id in np.unique(clusters):\n",
        "    cluster_points = result_df_with_clusters[clusters == cluster_id]\n",
        "    cluster_mean = np.mean(cluster_points, axis=0)\n",
        "    cluster_means.append(cluster_mean)\n",
        "\n",
        "nearest_nodes = []\n",
        "for cluster_id, cluster_mean in zip(np.unique(clusters), cluster_means):\n",
        "    cluster_points = result_df_with_clusters[clusters == cluster_id]\n",
        "    nearest_node_index = np.argmin(np.linalg.norm(cluster_points - cluster_mean, axis=1))\n",
        "    nearest_node = cluster_points.iloc[nearest_node_index]\n",
        "    nearest_nodes.append(nearest_node)"
      ],
      "metadata": {
        "id": "6v4NzB7u7KJH"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "pairwise = []\n",
        "stand_country_dict = {}\n",
        "for idx, series in enumerate(nearest_nodes):\n",
        "    nearest_node_id = series.name\n",
        "    stand_country_dict[idx+1] = result_df.iloc[nearest_node_id][0]\n",
        "    for point_id in result_df_with_clusters[clusters == idx+1].index:\n",
        "        target_lat = result_df.iloc[point_id]['latitude']\n",
        "        target_lng = result_df.iloc[point_id]['longitude']\n",
        "        source_lat = result_df.iloc[nearest_node_id]['latitude']\n",
        "        source_lng = result_df.iloc[nearest_node_id]['longitude']\n",
        "        dist = euclidean(result_df_with_clusters.iloc[point_id], result_df_with_clusters.iloc[nearest_node_id])\n",
        "        pairwise.append([idx+1, nearest_node_id,[source_lng,source_lat], point_id, [target_lng,target_lat],dist])\n",
        "link_df = pd.DataFrame(pairwise, columns = ['cluster','source','source_coordinates','target','target_coordinates','relationship'])\n",
        "node_df = pd.concat([extracted_prevalence_df, result_df_with_clusters[['hcluster']]], axis=1)\n",
        "node_df['coordinates'] = node_df.apply(lambda row: [row['longitude'],row['latitude']], axis=1)\n",
        "node_df = node_df[['state','prevalence_percent','hcluster','coordinates']]\n",
        "node_df['id'] = node_df.index"
      ],
      "metadata": {
        "id": "E6Nk-O_GA441"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "mean_pp = dict(node_df.groupby('hcluster')['prevalence_percent'].mean())\n",
        "len_memeber = dict(node_df.groupby('hcluster')['id'].count())\n",
        "\n",
        "mean_prevalence = []\n",
        "stand_country_name = []\n",
        "c_len_member = []\n",
        "node_df[node_df['id'] == 17]['state']\n",
        "for i in range(len(node_df)):\n",
        "    mean_prevalence.append(mean_pp[node_df.iloc[i]['hcluster']])\n",
        "    stand_country_name.append(stand_country_dict[node_df.iloc[i]['hcluster']])\n",
        "    c_len_member.append(len_memeber[node_df.iloc[i]['hcluster']])\n",
        "node_df = pd.concat([node_df, pd.DataFrame(mean_prevalence), pd.DataFrame(stand_country_name), pd.DataFrame(c_len_member)], axis=1)"
      ],
      "metadata": {
        "id": "fkIKCnsgkMBb"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "node_df.columns = ['state','prevalence_percent','hcluster','coordinates','id','mean_pervalence','representative','cluster_mem_number']\n",
        "# node_df"
      ],
      "metadata": {
        "id": "TRCf9IeumMcZ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "folder_path = '/content/drive/Shareddrives/DSCI554 project/Data collection/mental_health_data'\n",
        "abbrv_df = pd.read_csv(f'{folder_path}/abbrv.csv')\n",
        "node_df = pd.merge(node_df,abbrv_df,on='state',how='outer')\n",
        "# node_df"
      ],
      "metadata": {
        "id": "SsECQAETOYx8"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "d3_network = {}\n",
        "d3_network['nodes'] = node_df.to_dict(orient='records')\n",
        "d3_network['links'] = link_df.to_dict(orient='records')\n",
        "\n",
        "json_data = json.dumps(d3_network)\n",
        "\n",
        "with open('network_data.json', 'w') as json_file:\n",
        "    json_file.write(json_data)"
      ],
      "metadata": {
        "id": "d2yNDkugEGZT"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "min(link_df['relationship'])"
      ],
      "metadata": {
        "id": "R0IMoNHXVkE6"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "grouped bar chart - factors behind"
      ],
      "metadata": {
        "id": "T-fo400VLncO"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "factor_df = pd.read_csv(f'{folder_path}/factor.csv')\n",
        "factor_df = pd.merge(factor_df, node_df[['state','prevalence_percent','hcluster']],on='state',how='outer')"
      ],
      "metadata": {
        "id": "e5NCdQkTLt4I"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "scaler = StandardScaler()\n",
        "scaled_factor = scaler.fit_transform(factor_df.iloc[:,1:9])\n",
        "# len(scaled_factor[4 == factor_df['hcluster']])"
      ],
      "metadata": {
        "id": "961thH1KQq2c"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "corr_df = pd.DataFrame()\n",
        "for cluster_id in factor_df['hcluster'].unique():\n",
        "    cluster_scaled_factor = pd.DataFrame(scaled_factor[cluster_id == factor_df['hcluster']])\n",
        "    cluster_scaled_factor.columns = ['disability','unemployment','higher_education','uninsured','family_median_income','GDP','ACPSA','prevalence_percent']\n",
        "    cluster_corr = cluster_scaled_factor.corr()\n",
        "    corr_df = pd.concat([corr_df, pd.DataFrame(cluster_corr['prevalence_percent'])], axis=1)\n",
        "\n",
        "cluster_scaled_factor = pd.DataFrame(scaled_factor)\n",
        "cluster_scaled_factor.columns = ['disability','unemployment','higher_education','uninsured','family_median_income','GDP','ACPSA','prevalence_percent']\n",
        "cluster_corr = cluster_scaled_factor.corr()\n",
        "corr_df = pd.concat([corr_df, pd.DataFrame(cluster_corr['prevalence_percent'])], axis=1)\n",
        "corr_df.columns = list(str(v) for v in factor_df['hcluster'].unique()) + ['total']\n",
        "corr_df = corr_df.drop(columns = ['4'])\n",
        "corr_df = corr_df.drop('prevalence_percent')\n",
        "corr_df['name'] = ['disability rate','unemployment rate','higher education','healthcare uninsured','family income','state GDP','art and culture economic']\n",
        "corr_dict = corr_df.to_dict(orient='records')\n",
        "\n",
        "json_data = json.dumps(corr_dict)\n",
        "\n",
        "with open('factor_corr.json', 'w') as json_file:\n",
        "    json_file.write(json_data)\n"
      ],
      "metadata": {
        "id": "zfPPjvcmNOYT"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "heatmap - health_pattern_by_age"
      ],
      "metadata": {
        "id": "a53bzu1L17wW"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "folder_path = '/content/drive/Shareddrives/DSCI554 project/Data collection/mental_health_data/data1_MHA_dashboard'\n",
        "filenames = os.listdir(folder_path)\n",
        "age_df = pd.read_csv(f'{folder_path}/depression_age_df.csv')\n",
        "age_df = age_df[['state','latitude','longitude']]\n",
        "for filename in filenames:\n",
        "    if '_age_' in filename:\n",
        "        df = pd.read_csv(os.path.join(folder_path, filename))\n",
        "        for col in df.columns:\n",
        "            if '100K' in col and 'all_years' not in col and 'all_ages' not in col and 'trauma' not in col:\n",
        "                age_df = pd.merge(age_df,df[['state',f'{col}']],on=\"state\",how='outer')\n"
      ],
      "metadata": {
        "id": "rFjB7jtTZwQB"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "result_age_df = pd.DataFrame(columns = ['state','value','year','label','issue'])\n",
        "\n",
        "for col_id, col in enumerate(age_df.columns):\n",
        "    if col_id not in [0,1,2]:\n",
        "        age_df_temp = age_df[['state',col]]\n",
        "        year = col.split('-')[0]\n",
        "        label = col.split('-')[1]\n",
        "        issue = col.split('-')[2]\n",
        "        age_df_temp = pd.concat([age_df_temp,pd.DataFrame([year]*len(age_df_temp)),\n",
        "                                pd.DataFrame([label]*len(age_df_temp)),\n",
        "                                pd.DataFrame([issue]*len(age_df_temp))],axis=1)\n",
        "        age_df_temp.columns = ['state','value','year','label','issue']\n",
        "        result_age_df = pd.concat([result_age_df,age_df_temp],ignore_index=True)"
      ],
      "metadata": {
        "id": "fABuZnU-A0kL"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "age_dict = {}\n",
        "for state in result_age_df['state'].unique():\n",
        "    age_dict[state] = result_age_df[result_age_df['state']==state].to_dict(orient = 'records')\n",
        "\n",
        "result_mean_age_df = pd.DataFrame(columns = ['value','year','label','issue'])\n",
        "mean_age_df = pd.DataFrame(age_df.mean()).transpose()\n",
        "for col_id, col in enumerate(mean_age_df.columns):\n",
        "    if col_id not in [0,1]:\n",
        "        age_df_temp = mean_age_df[[col]]\n",
        "        year = col.split('-')[0]\n",
        "        label = col.split('-')[1]\n",
        "        issue = col.split('-')[2]\n",
        "        age_df_temp = pd.concat([age_df_temp,pd.DataFrame([year]*len(age_df_temp)),\n",
        "                                pd.DataFrame([label]*len(age_df_temp)),\n",
        "                                pd.DataFrame([issue]*len(age_df_temp))],axis=1)\n",
        "        age_df_temp.columns = ['value','year','label','issue']\n",
        "        result_mean_age_df = pd.concat([result_mean_age_df,age_df_temp],ignore_index=True)\n",
        "\n",
        "age_dict['mean'] = result_mean_age_df.to_dict(orient='records')"
      ],
      "metadata": {
        "id": "UPPjFMfZEIPj"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "json_data = json.dumps(age_dict)\n",
        "\n",
        "with open('age_year.json', 'w') as json_file:\n",
        "    json_file.write(json_data)"
      ],
      "metadata": {
        "id": "1iy7XJXZH32o"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "state_prevalence"
      ],
      "metadata": {
        "id": "e7z0lOJQeX_U"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "state_df = pd.merge(result_df[['state','depression','suicide','ptsd','psychosis']], extracted_prevalence_df[['state','prevalence_percent']], on ='state',how='outer')"
      ],
      "metadata": {
        "id": "0ZXI5RBeeXNj"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "state_dict = state_df.to_dict(orient='records')\n",
        "json_data = json.dumps(state_dict)\n",
        "\n",
        "with open('state_prevalence.json', 'w') as json_file:\n",
        "    json_file.write(json_data)"
      ],
      "metadata": {
        "id": "T20p8OGnf60h"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "state_df.describe().transpose()[['count','mean','std','min','max']]"
      ],
      "metadata": {
        "id": "9BnEFmFvL248"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "heatmap_race"
      ],
      "metadata": {
        "id": "twQexhg8jwI4"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "folder_path = '/content/drive/Shareddrives/DSCI554 project/Data collection/mental_health_data/data1_MHA_dashboard'\n",
        "filenames = os.listdir(folder_path)\n",
        "races_df = pd.read_csv(f'{folder_path}/depression_age_df.csv')\n",
        "races_df = races_df[['state']]\n",
        "for filename in filenames:\n",
        "    if '_race_' in filename:\n",
        "        df = pd.read_csv(os.path.join(folder_path, filename))\n",
        "        for col in df.columns:\n",
        "            if '2023' in col and '100K' in col and (('american' in col) or ('asian' in col) or ('black' in col) or ('hispanic' in col) or ('more_than_one' in col)):\n",
        "                if ('ptsd' in col) or ('depression' in col) or ('suicide' in col) or ('psychosis' in col):\n",
        "                    races_df = pd.merge(races_df,df[['state',f'{col}']],on=\"state\",how='outer')\n",
        "races_df.fillna(races_df.mean(numeric_only=True), inplace=True)"
      ],
      "metadata": {
        "id": "Isw2mzJ-jvce"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "races_name = {\"american_indian_alaska_native\":\"American Indian/Alaska Native\",\n",
        "              \"asian_pacific_islander\":\"Asian/Pacific Islander\",\n",
        "              \"black_african_american\":\"Black/African American\",\n",
        "              \"hispanic_latino\":\"Hispanic/Latino\",\n",
        "              \"more_than_one\":\"Multiracial\"}"
      ],
      "metadata": {
        "id": "NjmKwim_uDYG"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "result_races_df = pd.DataFrame(columns = ['state','value','races','issue'])\n",
        "\n",
        "for col_id, col in enumerate(races_df.columns):\n",
        "    if col_id not in [0]:\n",
        "        races_df_temp = races_df[['state',col]]\n",
        "        races = races_name[col.split('-')[1]]\n",
        "        issue = col.split('-')[2]\n",
        "        races_df_temp = pd.concat([races_df_temp,\n",
        "                                pd.DataFrame([races]*len(races_df_temp)),\n",
        "                                pd.DataFrame([issue]*len(races_df_temp))],axis=1)\n",
        "        races_df_temp.columns =['state','value','races','issue']\n",
        "        result_races_df = pd.concat([result_races_df,races_df_temp],ignore_index=True)"
      ],
      "metadata": {
        "id": "rURMQGNvtWc-"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "races_dict = {}\n",
        "for state in result_races_df['state'].unique():\n",
        "    races_dict[state] = result_races_df[result_races_df['state']==state].to_dict(orient = 'records')\n"
      ],
      "metadata": {
        "id": "UaKDD-B-uvm1"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "\n",
        "result_mean_races_df = pd.DataFrame(columns =['value','races','issue'])\n",
        "mean_races_df = pd.DataFrame(races_df.mean()).transpose()\n",
        "\n",
        "for col_id, col in enumerate(mean_races_df.columns):\n",
        "    if col_id not in []:\n",
        "        races_df_temp = mean_races_df[[col]]\n",
        "        races = races_name[col.split('-')[1]]\n",
        "        issue = col.split('-')[2]\n",
        "        races_df_temp = pd.concat([races_df_temp,\n",
        "                                pd.DataFrame([races]*len(races_df_temp)),\n",
        "                                pd.DataFrame([issue]*len(races_df_temp))],axis=1)\n",
        "        races_df_temp.columns = ['value','races','issue']\n",
        "        result_mean_races_df = pd.concat([result_mean_races_df,races_df_temp],ignore_index=True)\n",
        "\n",
        "races_dict['mean'] = result_mean_races_df.to_dict(orient='records')"
      ],
      "metadata": {
        "id": "-C9A_DEdwoD1"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "json_data = json.dumps(races_dict)\n",
        "\n",
        "with open('races_2023.json', 'w') as json_file:\n",
        "    json_file.write(json_data)"
      ],
      "metadata": {
        "id": "UcPcj4O-x27N"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}