Notebook_4_citation_network.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# CORD-19 citation network analysis\n",
    "\n",
    "In this notebook, we build a bibliographic coupling citation network of the publications we could match in Dimensions, and perform clustering using the Leiden algorithm. We export cluster partition information, as well as global and local (cluster-specific) centrality measures to use in the regression analysis."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The autoreload extension is already loaded. To reload it, use:\n",
      "  %reload_ext autoreload\n"
     ]
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "# magics and warnings\n",
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "import warnings; warnings.simplefilter('ignore')\n",
    "\n",
    "import os, random, codecs, time, pickle\n",
    "import ujson as json\n",
    "from tqdm.notebook import tqdm\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "\n",
    "seed = 99\n",
    "random.seed(seed)\n",
    "np.random.seed(seed)\n",
    "\n",
    "import nltk, sklearn\n",
    "import matplotlib.pyplot as plt\n",
    "from  matplotlib.ticker import FuncFormatter\n",
    "import seaborn as sns\n",
    "sns.set(style=\"white\")\n",
    "sns.set_context(\"paper\", font_scale=1.2, rc={\"lines.linewidth\": 2.5})"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# dataset Dimensions: source of citation data\n",
    "\n",
    "json_sources = \"../cwts_covid/datasets_output/json_dimensions_cwts\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "# build networks and node metadata\n",
    "# this requires Dimensions data, downloaded following: https://github.com/CWTSLeiden/cwts_covid\n",
    "\n",
    "from collections import OrderedDict\n",
    "import sys\n",
    "\n",
    "metadata = list()\n",
    "metadata_dict = dict()\n",
    "citations = OrderedDict()\n",
    "total_publications = 0\n",
    "\n",
    "for root, dirs, files in os.walk(json_sources):\n",
    "    for f in files:\n",
    "        try:\n",
    "            all_data = codecs.open(os.path.join(root,f)).read()\n",
    "        except:\n",
    "            continue\n",
    "        for line in all_data.split(\"\\n\"):\n",
    "            total_publications += 1\n",
    "            if line:\n",
    "                try:\n",
    "                    data = json.loads(line)\n",
    "                    if \"references\" in data.keys() and len(data[\"references\"])>0: # filter out papers without outbound citations\n",
    "                        title = \"\"\n",
    "                        year = \"\"\n",
    "                        doi = \"\"\n",
    "                        pmid = \"\"\n",
    "                        d_id = \"\"\n",
    "                        times_cited = \"\"\n",
    "                        journal = \"\"\n",
    "                        if \"title\" in data.keys() and data[\"title\"]:\n",
    "                            title = data[\"title\"].replace('\"',\"\")\n",
    "                        if \"journal\" in data.keys() and \"title\" in data[\"journal\"].keys() and data[\"journal\"][\"title\"]:\n",
    "                            jorunal = data[\"journal\"][\"title\"]\n",
    "                        if \"year\" in data.keys() and data[\"year\"]:\n",
    "                            year = str(data[\"year\"])\n",
    "                        if \"doi\" in data.keys() and data[\"doi\"]:\n",
    "                            doi = data[\"doi\"]\n",
    "                        if \"pmid\" in data.keys() and data[\"pmid\"]:\n",
    "                            doi = str(data[\"pmid\"])\n",
    "                        if \"id\" in data.keys() and data[\"id\"]:\n",
    "                            d_id = str(data[\"id\"])\n",
    "                        if \"times_cited\" in data.keys():\n",
    "                            times_cited = str(data[\"times_cited\"])\n",
    "                        metadata.append((d_id,'\"'+title+'\"',year,'\"'+doi+'\"',pmid,'\"'+journal+'\"',times_cited))\n",
    "                        metadata_dict[d_id] = {\"title\":title,\"year\":year,\"doi\":doi,\"journal\":journal,\"times_cited\":times_cited,\n",
    "                                                    \"pmid\":pmid,\"d_id\":d_id}\n",
    "                        citations[data[\"id\"]] = data[\"references\"]\n",
    "                except ValueError:\n",
    "                    print(\"Could not convert data.\")\n",
    "                except:\n",
    "                    print(line)\n",
    "                    continue"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "118214"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(citations)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "150765"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "total_publications"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "fe48e5739b3c4c18aabaf5e7c69501e9",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "HBox(children=(FloatProgress(value=1.0, bar_style='info', max=1.0), HTML(value='')))"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n"
     ]
    }
   ],
   "source": [
    "# create bibc network\n",
    "# this is a super naive method, and takes a while\n",
    "\n",
    "bibc_edge_list = list()\n",
    "items = [(k,v) for k,v in citations.items()]\n",
    "\n",
    "for n,(k1,v1) in tqdm(enumerate(items)):\n",
    "    for k2,v2 in items[n+1:]:\n",
    "        edge_weight = len(set(v2).intersection(set(v1))) / len(set(v2).union(set(v1))) # fractional edge weight\n",
    "        if edge_weight > 0:\n",
    "            bibc_edge_list.append((k1,k2,str(edge_weight)))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# export in csv (optional)\n",
    "\n",
    "with open(\"data/bib_coupling_edgelist_norm.csv\",\"w\") as f:\n",
    "    f.write(\"Source,Target,Weight\\n\")\n",
    "    for edge in bibc_edge_list:\n",
    "        #if float(edge[2])>1:\n",
    "        f.write(\",\".join(edge)+\"\\n\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open(\"data/bib_coupling_nodelist.csv\",\"w\") as f:\n",
    "    f.write(\"ID,Title,Year,DOI,PMID,Journal,Times_cited\\n\")\n",
    "    for node in metadata:\n",
    "        f.write(\",\".join(node)+\"\\n\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "70091752"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(bibc_edge_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "weight_list = [float(x[2]) for x in bibc_edge_list if float(x[2])>0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.021739130434782608"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.median(weight_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1cc744ee50>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.distplot(weight_list)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "# create an igraph graph\n",
    "\n",
    "import igraph as ig\n",
    "ig.set_random_number_generator(random.seed(seed))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "g = ig.Graph(directed=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "edges = [(int(k[0].replace(\"pub.\",\"\")),int(k[1].replace(\"pub.\",\"\"))) for k in bibc_edge_list]\n",
    "nodes = list(set([k[0] for k in edges]).union(set([k[1] for k in edges])))\n",
    "weights = [float(k[2]) for k in bibc_edge_list]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [],
   "source": [
    "# mapping of nodes to identifiers\n",
    "\n",
    "node_map = {k:n for n,k in enumerate(nodes)}\n",
    "edges_int = [(node_map[k[0]],node_map[k[1]]) for k in edges]\n",
    "nodes_int = [node_map[k] for k in nodes]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "115116"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(node_map)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'There are 3098 disconnected publications (with references)'"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\"There are \"+str(len(citations)-len(node_map))+\" disconnected publications (with references)\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [],
   "source": [
    "inverse_node_map = {n:k for k,n in node_map.items()}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [],
   "source": [
    "# add nodes and edges\n",
    "\n",
    "g.add_vertices(nodes_int)\n",
    "g.add_edges(edges_int)\n",
    "g.es[\"weight\"] = weights\n",
    "g.vs[\"weight\"] = 1\n",
    "g.vs[\"dimensions_id\"] = nodes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [],
   "source": [
    "# clustering \n",
    "import leidenalg as la"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "h = g.components(mode='weak').giant()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "114829"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(h.vs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "# 0.05 seems to find a relatively high modularity and reasonable sized clusters. Lowering it brings up one or two large clusters.\n",
    "\n",
    "partition = la.find_partition(h, la.CPMVertexPartition, resolution_parameter = 0.05)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "86016459.00001101"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "partition.quality()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.34600157440217927"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "partition.q"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# distribution of the size of clusters\n",
    "\n",
    "b = sns.distplot([x for x in partition.membership if x<100], norm_hist=False, kde=False, bins=100)\n",
    "b.set_xlabel(\"Cluster\",fontsize=15)\n",
    "b.set_ylabel(\"Size\",fontsize=15)\n",
    "b.tick_params(labelsize=12)\n",
    "plt.tight_layout()\n",
    "plt.savefig(\"figures/nb4_cluster_size_distr.pdf\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "from collections import Counter\n",
    "\n",
    "c = Counter(partition.membership)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "6187"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(c)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[(0, 15749),\n",
       " (1, 8034),\n",
       " (2, 3205),\n",
       " (3, 1816),\n",
       " (4, 1669),\n",
       " (5, 1619),\n",
       " (6, 1522),\n",
       " (7, 1418),\n",
       " (8, 1344),\n",
       " (9, 1153),\n",
       " (10, 1083),\n",
       " (11, 1042),\n",
       " (12, 1014),\n",
       " (13, 1008),\n",
       " (14, 981)]"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "c.most_common(15)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "cluster_sizes = {c:s for c,s in c.most_common()}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "# add note attributes and calculate indicators. This takes some time to compute"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "h.vs['cluster_id'] = partition.membership\n",
    "h.vs['cluster_above_threshold'] = False\n",
    "h.vs['cluster_size'] = [cluster_sizes[x] for x in partition.membership]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "h.vs['global_degree'] = h.degree()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "h.vs['global_pagerank'] = h.pagerank(directed=False, weights='weight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "# using a cutoff here\n",
    "#h.vs['global_closeness'] = h.closeness(weights='weight')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "# cluster-specific indicators\n",
    "h.vs['local_degree'] = np.nan\n",
    "h.vs['local_pagerank'] = np.nan\n",
    "#h.vs['local_closeness'] = np.nan"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.507075738707121"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "threshold = 300 # we only calculate indicators for nodes in clusters of size equal of greater than this threshold\n",
    "\n",
    "# we pick a treshold to consider > 50% of the nodes\n",
    "sum([x[1] for x in c.most_common() if x[1]>=threshold])/len(h.vs)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "for cluster in c.most_common():\n",
    "    c_id = cluster[0]\n",
    "    c_size = cluster[1]\n",
    "    if c_size >= threshold:\n",
    "        t = partition.subgraph(c_id)\n",
    "        t.vs['local_degree'] = t.degree()\n",
    "        t.vs['local_pagerank'] = t.pagerank(directed=False, weights='weight')\n",
    "        for v in t.vs:\n",
    "            h.vs.find(dimensions_id=v['dimensions_id'])['local_degree'] = v['local_degree']\n",
    "            h.vs.find(dimensions_id=v['dimensions_id'])['local_pagerank'] = v['local_pagerank']\n",
    "            h.vs.find(dimensions_id=v['dimensions_id'])['cluster_above_threshold'] = True"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "# create and export a data frame\n",
    "df_network = pd.DataFrame.from_dict({\"dimensions_id\": [\"pub.\"+str(i) for i in h.vs['dimensions_id']],\n",
    "                                     \"network_cluster_id\": h.vs['cluster_id'],\n",
    "                                     \"network_cluster_above_threshold\": h.vs['cluster_above_threshold'],\n",
    "                                     \"network_cluster_size\": h.vs['cluster_size'],\n",
    "                                     \"network_global_degree\": h.vs['global_degree'],\n",
    "                                     \"network_global_pagerank\": h.vs['global_pagerank'],\n",
    "                                     \"network_local_degree\": h.vs['local_degree'],\n",
    "                                     \"network_local_pagerank\": h.vs['local_pagerank']})\n",
    "df_network.to_csv(\"data/df_network.csv\", index=False, compression=\"gzip\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "# graph of clusters (optional)\n",
    "\n",
    "H = partition.cluster_graph(combine_vertices={'weight': 'sum'},\n",
    "                                                 combine_edges={'weight': 'mean'})\n",
    "H.write(\"data/network_of_clusters.graphml\",format=\"graphml\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Plots"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "# you can select here if to use an LDA or a CTM topic model\n",
    "df_topic_intensity = pd.read_csv(\"data/df_topic_intensity.csv\", compression=\"gzip\")\n",
    "#df_topic_intensity = pd.read_csv(\"data/df_topic_intensity_ctm.csv\", compression=\"gzip\")\n",
    "df_joined_all = pd.read_csv(\"data/df_joined_all.csv\", compression=\"gzip\")\n",
    "df_network = pd.read_csv(\"data/df_network.csv\", compression=\"gzip\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>tm_coronaviruses</th>\n",
       "      <th>tm_epidemics</th>\n",
       "      <th>tm_ph</th>\n",
       "      <th>tm_mbi</th>\n",
       "      <th>tm_clinical_medicine</th>\n",
       "      <th>pub_id</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.335269</td>\n",
       "      <td>0.091129</td>\n",
       "      <td>0.111784</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.457265</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.299387</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.198364</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.422208</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.953206</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.141524</td>\n",
       "      <td>0.674861</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.165509</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.676764</td>\n",
       "      <td>0.307801</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   tm_coronaviruses  tm_epidemics     tm_ph  tm_mbi  tm_clinical_medicine  \\\n",
       "0          0.335269      0.091129  0.111784     0.0              0.457265   \n",
       "1          0.299387      0.000000  0.198364     0.0              0.422208   \n",
       "2          0.000000      0.000000  0.953206     0.0              0.000000   \n",
       "3          0.000000      0.141524  0.674861     0.0              0.165509   \n",
       "4          0.000000      0.676764  0.307801     0.0              0.000000   \n",
       "\n",
       "   pub_id  \n",
       "0       0  \n",
       "1       1  \n",
       "2       2  \n",
       "3       3  \n",
       "4       4  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_topic_intensity.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['tm_coronaviruses', 'tm_epidemics', 'tm_ph', 'tm_mbi',\n",
       "       'tm_clinical_medicine', 'pub_id'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_topic_intensity.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "topic_labels = ['tm_coronaviruses', 'tm_epidemics', 'tm_ph', 'tm_mbi',\n",
    "       'tm_clinical_medicine']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "tm_names = {\"Coronaviruses\": \"tm_coronaviruses\",\n",
    "    \"Epidemics\": \"tm_epidemics\",\n",
    "    \"Public health\": \"tm_ph\",\n",
    "    \"Molecular biology and immunology\": \"tm_mbi\",\n",
    "    \"Clinical medicine\": \"tm_clinical_medicine\"}\n",
    "topic_labels = [x for x in tm_names.keys()]\n",
    "reverse_tm_names = {x:y for y,x in tm_names.items()}\n",
    "df_topic_intensity.rename(columns=reverse_tm_names,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_embeddings = df_network.merge(df_joined_all[[\"pub_id\",\"dimensions_id\",\"in_wikipedia\",\"title\",\"publication_year\"]], how=\"inner\", left_on=\"dimensions_id\", right_on=\"dimensions_id\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(108213, 12)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_embeddings.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_embeddings = df_embeddings.merge(df_topic_intensity, how=\"inner\", left_on=\"pub_id\", right_on=\"pub_id\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(152247, 6)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_topic_intensity.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(105915, 17)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_embeddings.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>dimensions_id</th>\n",
       "      <th>network_cluster_id</th>\n",
       "      <th>network_cluster_above_threshold</th>\n",
       "      <th>network_cluster_size</th>\n",
       "      <th>network_global_degree</th>\n",
       "      <th>network_global_pagerank</th>\n",
       "      <th>network_local_degree</th>\n",
       "      <th>network_local_pagerank</th>\n",
       "      <th>pub_id</th>\n",
       "      <th>in_wikipedia</th>\n",
       "      <th>title</th>\n",
       "      <th>publication_year</th>\n",
       "      <th>Coronaviruses</th>\n",
       "      <th>Epidemics</th>\n",
       "      <th>Public health</th>\n",
       "      <th>Molecular biology and immunology</th>\n",
       "      <th>Clinical medicine</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>pub.1121976320</td>\n",
       "      <td>1075</td>\n",
       "      <td>False</td>\n",
       "      <td>12</td>\n",
       "      <td>60</td>\n",
       "      <td>0.000003</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>137626</td>\n",
       "      <td>False</td>\n",
       "      <td>An Exploration of Machine Learning Methods for...</td>\n",
       "      <td>2019.0</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.328873</td>\n",
       "      <td>0.397463</td>\n",
       "      <td>0.037389</td>\n",
       "      <td>0.232290</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>pub.1103626240</td>\n",
       "      <td>36</td>\n",
       "      <td>True</td>\n",
       "      <td>418</td>\n",
       "      <td>320</td>\n",
       "      <td>0.000005</td>\n",
       "      <td>136.0</td>\n",
       "      <td>0.002279</td>\n",
       "      <td>65622</td>\n",
       "      <td>False</td>\n",
       "      <td>Molecularly specific detection of bacterial li...</td>\n",
       "      <td>2018.0</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.069453</td>\n",
       "      <td>0.093479</td>\n",
       "      <td>0.506221</td>\n",
       "      <td>0.321498</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>pub.1049624581</td>\n",
       "      <td>23</td>\n",
       "      <td>True</td>\n",
       "      <td>655</td>\n",
       "      <td>48</td>\n",
       "      <td>0.000005</td>\n",
       "      <td>40.0</td>\n",
       "      <td>0.000637</td>\n",
       "      <td>103417</td>\n",
       "      <td>False</td>\n",
       "      <td>Hemodynamic-morphologic discriminants for intr...</td>\n",
       "      <td>2011.0</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.542807</td>\n",
       "      <td>0.064618</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.389494</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>pub.1026031634</td>\n",
       "      <td>2</td>\n",
       "      <td>True</td>\n",
       "      <td>3205</td>\n",
       "      <td>839</td>\n",
       "      <td>0.000009</td>\n",
       "      <td>567.0</td>\n",
       "      <td>0.000303</td>\n",
       "      <td>116401</td>\n",
       "      <td>False</td>\n",
       "      <td>Clinical Presentation and Birth Outcomes Assoc...</td>\n",
       "      <td>2016.0</td>\n",
       "      <td>0.48604</td>\n",
       "      <td>0.056072</td>\n",
       "      <td>0.037600</td>\n",
       "      <td>0.019100</td>\n",
       "      <td>0.397733</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>pub.1124597780</td>\n",
       "      <td>48</td>\n",
       "      <td>False</td>\n",
       "      <td>275</td>\n",
       "      <td>224</td>\n",
       "      <td>0.000003</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>111089</td>\n",
       "      <td>False</td>\n",
       "      <td>PfSWIB, a potential chromatin regulator for va...</td>\n",
       "      <td>2020.0</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.196043</td>\n",
       "      <td>0.035244</td>\n",
       "      <td>0.717387</td>\n",
       "      <td>0.040151</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "    dimensions_id  network_cluster_id  network_cluster_above_threshold  \\\n",
       "0  pub.1121976320                1075                            False   \n",
       "1  pub.1103626240                  36                             True   \n",
       "2  pub.1049624581                  23                             True   \n",
       "3  pub.1026031634                   2                             True   \n",
       "4  pub.1124597780                  48                            False   \n",
       "\n",
       "   network_cluster_size  network_global_degree  network_global_pagerank  \\\n",
       "0                    12                     60                 0.000003   \n",
       "1                   418                    320                 0.000005   \n",
       "2                   655                     48                 0.000005   \n",
       "3                  3205                    839                 0.000009   \n",
       "4                   275                    224                 0.000003   \n",
       "\n",
       "   network_local_degree  network_local_pagerank  pub_id  in_wikipedia  \\\n",
       "0                   NaN                     NaN  137626         False   \n",
       "1                 136.0                0.002279   65622         False   \n",
       "2                  40.0                0.000637  103417         False   \n",
       "3                 567.0                0.000303  116401         False   \n",
       "4                   NaN                     NaN  111089         False   \n",
       "\n",
       "                                               title  publication_year  \\\n",
       "0  An Exploration of Machine Learning Methods for...            2019.0   \n",
       "1  Molecularly specific detection of bacterial li...            2018.0   \n",
       "2  Hemodynamic-morphologic discriminants for intr...            2011.0   \n",
       "3  Clinical Presentation and Birth Outcomes Assoc...            2016.0   \n",
       "4  PfSWIB, a potential chromatin regulator for va...            2020.0   \n",
       "\n",
       "   Coronaviruses  Epidemics  Public health  Molecular biology and immunology  \\\n",
       "0        0.00000   0.328873       0.397463                          0.037389   \n",
       "1        0.00000   0.069453       0.093479                          0.506221   \n",
       "2        0.00000   0.542807       0.064618                          0.000000   \n",
       "3        0.48604   0.056072       0.037600                          0.019100   \n",
       "4        0.00000   0.196043       0.035244                          0.717387   \n",
       "\n",
       "   Clinical medicine  \n",
       "0           0.232290  \n",
       "1           0.321498  \n",
       "2           0.389494  \n",
       "3           0.397733  \n",
       "4           0.040151  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_embeddings.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# distribution of centralities for papers cited from Wikipedia or not\n",
    "\n",
    "which_centrality = \"network_global_degree\"\n",
    "\n",
    "plt.figure(figsize=(10,8))\n",
    "sns.distplot(df_embeddings[df_embeddings.in_wikipedia==True][which_centrality].values, color=\"green\", kde=True, norm_hist=True, kde_kws={\"shade\": True}, hist=False, label=\"In Wikipedia\")\n",
    "sns.distplot(df_embeddings[df_embeddings.in_wikipedia==False][which_centrality].values, color=\"blue\", kde=True, norm_hist=True, kde_kws={\"shade\": True}, hist=False, label=\"Not in Wikipedia\")\n",
    "plt.legend(fontsize=15)\n",
    "plt.xlabel(\"%s\"%which_centrality,fontsize=15)\n",
    "plt.savefig(\"figures/nb4_%s_in_wiki.pdf\"%which_centrality)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [],
   "source": [
    "# plot cluster % of papers cited from Wikipedia, against cluster size"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "threshold = 300 #see above"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "cluster_coverage = list()\n",
    "cluster_size = list()\n",
    "cluster_number = list()\n",
    "below_threshold_size = 0\n",
    "below_threshold_coverage = 0\n",
    "local_threshold = threshold\n",
    "#threshold = 500\n",
    "\n",
    "# also calculate paper centralities wrt. cluster centroid\n",
    "# for every paper, this is 1 - the (cosine) distance wrt. its centroid\n",
    "# we need both similarity and args to index results\n",
    "\n",
    "for cl in list(set(df_embeddings.network_cluster_id.to_list())):\n",
    "    cl_size = df_embeddings[df_embeddings.network_cluster_id == cl].shape[0]\n",
    "    inw = df_embeddings[(df_embeddings.in_wikipedia==True)&(df_embeddings.network_cluster_id==cl)].shape[0]\n",
    "    if cl_size>=threshold:\n",
    "        coverage = inw/cl_size\n",
    "        cluster_coverage.append(coverage)\n",
    "        cluster_size.append(cl_size)\n",
    "        cluster_number.append(cl)\n",
    "    else:\n",
    "        below_threshold_coverage += inw\n",
    "        below_threshold_size += cl_size\n",
    "#cluster_coverage.append(below_threshold_coverage/below_threshold_size)\n",
    "#cluster_size.append(below_threshold_size)\n",
    "#cluster_number.append(-1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "41"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(cluster_coverage)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(6,6))\n",
    "pl = sns.scatterplot(x=cluster_size,y=cluster_coverage,s=200)\n",
    "# add annotations one by one with a loop\n",
    "for dot in cluster_number:\n",
    "     pl.text(cluster_size[dot]+7, cluster_coverage[dot]+0.002, s=cluster_number[dot], horizontalalignment='left', size='medium', color='black', weight='semibold')\n",
    "plt.ylim((0,0.11))\n",
    "plt.xlim((0,max(cluster_size)+500))\n",
    "#plt.legend(fontsize=15)\n",
    "plt.tick_params(labelsize=12)\n",
    "plt.ylabel(\"% cited from Wikipedia\",fontsize=15)\n",
    "plt.xlabel(\"Cluster size\",fontsize=15)\n",
    "plt.savefig(\"figures/nb4_cluster_size_vs_wiki_coverage_%d_no_max.pdf\"%local_threshold)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [],
   "source": [
    "which_cluster = 31"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.distplot(df_embeddings[df_embeddings.network_cluster_id == which_cluster].publication_year, bins=20, kde=False)\n",
    "plt.xlabel(\"Year\",fontsize=15)\n",
    "plt.ylabel(\"Count\",fontsize=15)\n",
    "plt.tick_params(labelsize=12)\n",
    "plt.gca().xaxis.set_major_formatter(FuncFormatter(lambda x, _: int(x)))\n",
    "plt.tight_layout()\n",
    "#plt.savefig(\"figures/nb4_cluster_%d_yeardistr.pdf\"%(which_cluster))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(8,5))\n",
    "df_sub = df_embeddings[df_embeddings.network_cluster_id == which_cluster][[mt for mt in topic_labels]].mean()\n",
    "b = sns.barplot(y=df_sub.index,x=df_sub.values)\n",
    "b.set_xlabel(\"Proportion\",fontsize=15)\n",
    "b.set_ylabel(\"Macrotopic\",fontsize=15)\n",
    "b.tick_params(labelsize=12)\n",
    "plt.tight_layout()\n",
    "#plt.savefig(\"figures/nb4_cluster_%d_to_topicmodel.pdf\"%(which_cluster))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>dimensions_id</th>\n",
       "      <th>network_cluster_id</th>\n",
       "      <th>network_cluster_size</th>\n",
       "      <th>network_global_degree</th>\n",
       "      <th>network_global_pagerank</th>\n",
       "      <th>title</th>\n",
       "      <th>publication_year</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>8962</th>\n",
       "      <td>pub.1033411825</td>\n",
       "      <td>31</td>\n",
       "      <td>462</td>\n",
       "      <td>1176</td>\n",
       "      <td>0.000013</td>\n",
       "      <td>Characterization of the frameshift signal of E...</td>\n",
       "      <td>2005.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>61987</th>\n",
       "      <td>pub.1020049797</td>\n",
       "      <td>31</td>\n",
       "      <td>462</td>\n",
       "      <td>879</td>\n",
       "      <td>0.000013</td>\n",
       "      <td>The role of RNA pseudoknot stem 1 length in th...</td>\n",
       "      <td>1999.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>101359</th>\n",
       "      <td>pub.1041739343</td>\n",
       "      <td>31</td>\n",
       "      <td>462</td>\n",
       "      <td>712</td>\n",
       "      <td>0.000013</td>\n",
       "      <td>Pseudoknot-Dependent Programmed —1 Ribosomal F...</td>\n",
       "      <td>2009.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100543</th>\n",
       "      <td>pub.1006608266</td>\n",
       "      <td>31</td>\n",
       "      <td>462</td>\n",
       "      <td>1010</td>\n",
       "      <td>0.000013</td>\n",
       "      <td>The stimulatory RNA of the Visna-Maedi retrovi...</td>\n",
       "      <td>2008.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49617</th>\n",
       "      <td>pub.1008977174</td>\n",
       "      <td>31</td>\n",
       "      <td>462</td>\n",
       "      <td>1553</td>\n",
       "      <td>0.000013</td>\n",
       "      <td>Transactivation of programmed ribosomal frames...</td>\n",
       "      <td>2014.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>100225</th>\n",
       "      <td>pub.1024956862</td>\n",
       "      <td>31</td>\n",
       "      <td>462</td>\n",
       "      <td>864</td>\n",
       "      <td>0.000012</td>\n",
       "      <td>A Genome-Wide Analysis of RNA Pseudoknots That...</td>\n",
       "      <td>2013.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>53562</th>\n",
       "      <td>pub.1000083772</td>\n",
       "      <td>31</td>\n",
       "      <td>462</td>\n",
       "      <td>826</td>\n",
       "      <td>0.000012</td>\n",
       "      <td>Identification and analysis of the gag-pol rib...</td>\n",
       "      <td>1992.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>62079</th>\n",
       "      <td>pub.1049934732</td>\n",
       "      <td>31</td>\n",
       "      <td>462</td>\n",
       "      <td>762</td>\n",
       "      <td>0.000011</td>\n",
       "      <td>V, 2.Ribosomal frameshifting in astroviruses</td>\n",
       "      <td>2004.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>102514</th>\n",
       "      <td>pub.1052229334</td>\n",
       "      <td>31</td>\n",
       "      <td>462</td>\n",
       "      <td>502</td>\n",
       "      <td>0.000011</td>\n",
       "      <td>An intermolecular RNA triplex provides insight...</td>\n",
       "      <td>2009.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>104115</th>\n",
       "      <td>pub.1028642862</td>\n",
       "      <td>31</td>\n",
       "      <td>462</td>\n",
       "      <td>895</td>\n",
       "      <td>0.000011</td>\n",
       "      <td>The many paths to frameshifting: kinetic model...</td>\n",
       "      <td>2010.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>103888</th>\n",
       "      <td>pub.1006621416</td>\n",
       "      <td>31</td>\n",
       "      <td>462</td>\n",
       "      <td>1095</td>\n",
       "      <td>0.000010</td>\n",
       "      <td>An Extended Signal Involved in Eukaryotic −1 F...</td>\n",
       "      <td>2005.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23614</th>\n",
       "      <td>pub.1012522136</td>\n",
       "      <td>31</td>\n",
       "      <td>462</td>\n",
       "      <td>726</td>\n",
       "      <td>0.000010</td>\n",
       "      <td>HIV-1 frameshift efficiency is primarily deter...</td>\n",
       "      <td>2012.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29792</th>\n",
       "      <td>pub.1004688837</td>\n",
       "      <td>31</td>\n",
       "      <td>462</td>\n",
       "      <td>960</td>\n",
       "      <td>0.000010</td>\n",
       "      <td>Possible involvement of coaxially stacked doub...</td>\n",
       "      <td>2015.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>52624</th>\n",
       "      <td>pub.1038351719</td>\n",
       "      <td>31</td>\n",
       "      <td>462</td>\n",
       "      <td>794</td>\n",
       "      <td>0.000010</td>\n",
       "      <td>Evidence for an RNA pseudoknot loop-helix inte...</td>\n",
       "      <td>1999.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2003</th>\n",
       "      <td>pub.1044390457</td>\n",
       "      <td>31</td>\n",
       "      <td>462</td>\n",
       "      <td>428</td>\n",
       "      <td>0.000010</td>\n",
       "      <td>Programmed -1 frameshifting efficiency correla...</td>\n",
       "      <td>2012.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>66487</th>\n",
       "      <td>pub.1040519453</td>\n",
       "      <td>31</td>\n",
       "      <td>462</td>\n",
       "      <td>664</td>\n",
       "      <td>0.000010</td>\n",
       "      <td>Prokaryotic-style frameshifting in a plant tra...</td>\n",
       "      <td>2003.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38065</th>\n",
       "      <td>pub.1090714842</td>\n",
       "      <td>31</td>\n",
       "      <td>462</td>\n",
       "      <td>854</td>\n",
       "      <td>0.000010</td>\n",
       "      <td>Determination of Ferret Enteric Coronavirus Ge...</td>\n",
       "      <td>2017.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>71506</th>\n",
       "      <td>pub.1006468765</td>\n",
       "      <td>31</td>\n",
       "      <td>462</td>\n",
       "      <td>622</td>\n",
       "      <td>0.000010</td>\n",
       "      <td>Identification of RNA pseudoknot-binding ligan...</td>\n",
       "      <td>2011.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50201</th>\n",
       "      <td>pub.1052495971</td>\n",
       "      <td>31</td>\n",
       "      <td>462</td>\n",
       "      <td>492</td>\n",
       "      <td>0.000010</td>\n",
       "      <td>Translating old drugs into new treatments: rib...</td>\n",
       "      <td>1998.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>62071</th>\n",
       "      <td>pub.1000651620</td>\n",
       "      <td>31</td>\n",
       "      <td>462</td>\n",
       "      <td>719</td>\n",
       "      <td>0.000009</td>\n",
       "      <td>Frameshifting RNA Pseudoknots: Structure and M...</td>\n",
       "      <td>2009.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         dimensions_id  network_cluster_id  network_cluster_size  \\\n",
       "8962    pub.1033411825                  31                   462   \n",
       "61987   pub.1020049797                  31                   462   \n",
       "101359  pub.1041739343                  31                   462   \n",
       "100543  pub.1006608266                  31                   462   \n",
       "49617   pub.1008977174                  31                   462   \n",
       "100225  pub.1024956862                  31                   462   \n",
       "53562   pub.1000083772                  31                   462   \n",
       "62079   pub.1049934732                  31                   462   \n",
       "102514  pub.1052229334                  31                   462   \n",
       "104115  pub.1028642862                  31                   462   \n",
       "103888  pub.1006621416                  31                   462   \n",
       "23614   pub.1012522136                  31                   462   \n",
       "29792   pub.1004688837                  31                   462   \n",
       "52624   pub.1038351719                  31                   462   \n",
       "2003    pub.1044390457                  31                   462   \n",
       "66487   pub.1040519453                  31                   462   \n",
       "38065   pub.1090714842                  31                   462   \n",
       "71506   pub.1006468765                  31                   462   \n",
       "50201   pub.1052495971                  31                   462   \n",
       "62071   pub.1000651620                  31                   462   \n",
       "\n",
       "        network_global_degree  network_global_pagerank  \\\n",
       "8962                     1176                 0.000013   \n",
       "61987                     879                 0.000013   \n",
       "101359                    712                 0.000013   \n",
       "100543                   1010                 0.000013   \n",
       "49617                    1553                 0.000013   \n",
       "100225                    864                 0.000012   \n",
       "53562                     826                 0.000012   \n",
       "62079                     762                 0.000011   \n",
       "102514                    502                 0.000011   \n",
       "104115                    895                 0.000011   \n",
       "103888                   1095                 0.000010   \n",
       "23614                     726                 0.000010   \n",
       "29792                     960                 0.000010   \n",
       "52624                     794                 0.000010   \n",
       "2003                      428                 0.000010   \n",
       "66487                     664                 0.000010   \n",
       "38065                     854                 0.000010   \n",
       "71506                     622                 0.000010   \n",
       "50201                     492                 0.000010   \n",
       "62071                     719                 0.000009   \n",
       "\n",
       "                                                    title  publication_year  \n",
       "8962    Characterization of the frameshift signal of E...            2005.0  \n",
       "61987   The role of RNA pseudoknot stem 1 length in th...            1999.0  \n",
       "101359  Pseudoknot-Dependent Programmed —1 Ribosomal F...            2009.0  \n",
       "100543  The stimulatory RNA of the Visna-Maedi retrovi...            2008.0  \n",
       "49617   Transactivation of programmed ribosomal frames...            2014.0  \n",
       "100225  A Genome-Wide Analysis of RNA Pseudoknots That...            2013.0  \n",
       "53562   Identification and analysis of the gag-pol rib...            1992.0  \n",
       "62079        V, 2.Ribosomal frameshifting in astroviruses            2004.0  \n",
       "102514  An intermolecular RNA triplex provides insight...            2009.0  \n",
       "104115  The many paths to frameshifting: kinetic model...            2010.0  \n",
       "103888  An Extended Signal Involved in Eukaryotic −1 F...            2005.0  \n",
       "23614   HIV-1 frameshift efficiency is primarily deter...            2012.0  \n",
       "29792   Possible involvement of coaxially stacked doub...            2015.0  \n",
       "52624   Evidence for an RNA pseudoknot loop-helix inte...            1999.0  \n",
       "2003    Programmed -1 frameshifting efficiency correla...            2012.0  \n",
       "66487   Prokaryotic-style frameshifting in a plant tra...            2003.0  \n",
       "38065   Determination of Ferret Enteric Coronavirus Ge...            2017.0  \n",
       "71506   Identification of RNA pseudoknot-binding ligan...            2011.0  \n",
       "50201   Translating old drugs into new treatments: rib...            1998.0  \n",
       "62071   Frameshifting RNA Pseudoknots: Structure and M...            2009.0  "
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# examples of central papers in a cluster\n",
    "\n",
    "df_embeddings[[\"dimensions_id\",\"network_cluster_id\",\"network_cluster_size\",\"network_global_degree\",\"network_global_pagerank\",\"title\",\"publication_year\"]][df_embeddings.network_cluster_id==which_cluster].sort_values(\"network_global_pagerank\",ascending=False).head(20)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}