data_gathering_preprocessing/sample_analysis_update.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Classifying Unpaywall Locations - Sample Analysis Update "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This notebook is a supplement and an update of the result of the last validation found under [Link](https://github.com/subugoe/oa-german-inst). Once again, an additional sample consisting of a hundred DOIs is examined. This time, a revised classification algorithm was used and I will therefore determine its effectiveness in comparison with the previous algorithm.\n",
    "\n",
    "The structure of this analysis follows the same pattern as last evaluation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "from matplotlib.ticker import MaxNLocator\n",
    "import seaborn as sns\n",
    "import numpy as np\n",
    "from sklearn.metrics import accuracy_score\n",
    "import networkx as nx"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "sns.set_style('whitegrid')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "df = pd.read_csv('sample_manual_check_2.csv', sep=';')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "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>upw_doi</th>\n",
       "      <th>oa_category</th>\n",
       "      <th>evidence</th>\n",
       "      <th>host_type</th>\n",
       "      <th>repository_institution</th>\n",
       "      <th>url_for_landing_page</th>\n",
       "      <th>url_for_pdf</th>\n",
       "      <th>evaluation</th>\n",
       "      <th>revision</th>\n",
       "      <th>comment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>10.1103/physrevd.89.054038</td>\n",
       "      <td>opendoar_subject</td>\n",
       "      <td>oa repository (via OAI-PMH doi match)</td>\n",
       "      <td>repository</td>\n",
       "      <td>Cornell University - arXiv</td>\n",
       "      <td>http://arxiv.org/abs/1312.5621</td>\n",
       "      <td>http://arxiv.org/pdf/1312.5621</td>\n",
       "      <td>True</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>10.1103/physrevd.89.054038</td>\n",
       "      <td>other_repo</td>\n",
       "      <td>oa repository (semantic scholar lookup)</td>\n",
       "      <td>repository</td>\n",
       "      <td>NaN</td>\n",
       "      <td>https://semanticscholar.org/paper/e9195a939f7c...</td>\n",
       "      <td>http://pdfs.semanticscholar.org/46b5/253209f4a...</td>\n",
       "      <td>True</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>10.1007/s15010-011-0155-z</td>\n",
       "      <td>not_oa</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>10.1051/0004-6361/201118194</td>\n",
       "      <td>other_oa_journal</td>\n",
       "      <td>open (via free pdf)</td>\n",
       "      <td>publisher</td>\n",
       "      <td>NaN</td>\n",
       "      <td>https://doi.org/10.1051/0004-6361/201118194</td>\n",
       "      <td>https://www.aanda.org/articles/aa/pdf/2012/04/...</td>\n",
       "      <td>True</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>10.1051/0004-6361/201118194</td>\n",
       "      <td>opendoar_subject</td>\n",
       "      <td>oa repository (via OAI-PMH doi match)</td>\n",
       "      <td>repository</td>\n",
       "      <td>arXiv.org</td>\n",
       "      <td>http://arxiv.org/abs/1110.0475</td>\n",
       "      <td>http://arxiv.org/pdf/1110.0475</td>\n",
       "      <td>True</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                       upw_doi       oa_category  \\\n",
       "0   10.1103/physrevd.89.054038  opendoar_subject   \n",
       "1   10.1103/physrevd.89.054038        other_repo   \n",
       "2    10.1007/s15010-011-0155-z            not_oa   \n",
       "3  10.1051/0004-6361/201118194  other_oa_journal   \n",
       "4  10.1051/0004-6361/201118194  opendoar_subject   \n",
       "\n",
       "                                  evidence   host_type  \\\n",
       "0    oa repository (via OAI-PMH doi match)  repository   \n",
       "1  oa repository (semantic scholar lookup)  repository   \n",
       "2                                      NaN         NaN   \n",
       "3                      open (via free pdf)   publisher   \n",
       "4    oa repository (via OAI-PMH doi match)  repository   \n",
       "\n",
       "       repository_institution  \\\n",
       "0  Cornell University - arXiv   \n",
       "1                         NaN   \n",
       "2                         NaN   \n",
       "3                         NaN   \n",
       "4                   arXiv.org   \n",
       "\n",
       "                                url_for_landing_page  \\\n",
       "0                     http://arxiv.org/abs/1312.5621   \n",
       "1  https://semanticscholar.org/paper/e9195a939f7c...   \n",
       "2                                                NaN   \n",
       "3        https://doi.org/10.1051/0004-6361/201118194   \n",
       "4                     http://arxiv.org/abs/1110.0475   \n",
       "\n",
       "                                         url_for_pdf  evaluation revision  \\\n",
       "0                     http://arxiv.org/pdf/1312.5621        True      NaN   \n",
       "1  http://pdfs.semanticscholar.org/46b5/253209f4a...        True      NaN   \n",
       "2                                                NaN        True      NaN   \n",
       "3  https://www.aanda.org/articles/aa/pdf/2012/04/...        True      NaN   \n",
       "4                     http://arxiv.org/pdf/1110.0475        True      NaN   \n",
       "\n",
       "  comment  \n",
       "0     NaN  \n",
       "1     NaN  \n",
       "2     NaN  \n",
       "3     NaN  \n",
       "4     NaN  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9456521739130435"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "accuracy_score([1] * len(df.evaluation), df.evaluation)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The accuracy of classifications performed on this dataset has increased by 5% compared to the result of last analysis. Overall, the accuracy is at 95%."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_false = df[df.evaluation == False]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "upw_doi\n",
       "10.1002/mma.3895              2\n",
       "10.1007/s00205-011-0464-5     1\n",
       "10.1039/c4nr07378a            2\n",
       "10.1103/physrevb.83.054109    1\n",
       "10.2961/jlmn.2014.03.0019     1\n",
       "10.3390/polym10090935         1\n",
       "10.3847/2041-8213/aa5fff      2\n",
       "Name: evaluation, dtype: int64"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_false.groupby(['upw_doi'])['evaluation'].count()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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>upw_doi</th>\n",
       "      <th>oa_category</th>\n",
       "      <th>evidence</th>\n",
       "      <th>host_type</th>\n",
       "      <th>repository_institution</th>\n",
       "      <th>url_for_landing_page</th>\n",
       "      <th>url_for_pdf</th>\n",
       "      <th>evaluation</th>\n",
       "      <th>revision</th>\n",
       "      <th>comment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>count</th>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>9</td>\n",
       "      <td>9</td>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>10</td>\n",
       "      <td>9</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>unique</th>\n",
       "      <td>7</td>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>9</td>\n",
       "      <td>10</td>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>top</th>\n",
       "      <td>10.1039/c4nr07378a</td>\n",
       "      <td>opendoar_other</td>\n",
       "      <td>oa repository (via OAI-PMH doi match)</td>\n",
       "      <td>repository</td>\n",
       "      <td>escidoc.org</td>\n",
       "      <td>http://bib-pubdb1.desy.de/search?p=id:%22PUBDB...</td>\n",
       "      <td>http://www.jlps.gr.jp/jlmn/upload/83602e721c8a...</td>\n",
       "      <td>False</td>\n",
       "      <td>opendoar_inst</td>\n",
       "      <td>https://v2.sherpa.ac.uk/id/repository/4170</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>freq</th>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>7</td>\n",
       "      <td>10</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>10</td>\n",
       "      <td>4</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                   upw_doi     oa_category  \\\n",
       "count                   10              10   \n",
       "unique                   7               3   \n",
       "top     10.1039/c4nr07378a  opendoar_other   \n",
       "freq                     2               5   \n",
       "\n",
       "                                     evidence   host_type  \\\n",
       "count                                      10          10   \n",
       "unique                                      2           1   \n",
       "top     oa repository (via OAI-PMH doi match)  repository   \n",
       "freq                                        7          10   \n",
       "\n",
       "       repository_institution  \\\n",
       "count                       9   \n",
       "unique                      5   \n",
       "top               escidoc.org   \n",
       "freq                        2   \n",
       "\n",
       "                                     url_for_landing_page  \\\n",
       "count                                                   9   \n",
       "unique                                                  9   \n",
       "top     http://bib-pubdb1.desy.de/search?p=id:%22PUBDB...   \n",
       "freq                                                    1   \n",
       "\n",
       "                                              url_for_pdf evaluation  \\\n",
       "count                                                  10         10   \n",
       "unique                                                 10          1   \n",
       "top     http://www.jlps.gr.jp/jlmn/upload/83602e721c8a...      False   \n",
       "freq                                                    1         10   \n",
       "\n",
       "             revision                                     comment  \n",
       "count              10                                           9  \n",
       "unique              5                                           5  \n",
       "top     opendoar_inst  https://v2.sherpa.ac.uk/id/repository/4170  \n",
       "freq                4                                           2  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_false.describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "About 5% of the locations have been classified incorrectly, thus the value has decreased by 5% compared to the previous result. Further, 7% of the articles have at least one faulty connection to a location."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax = df_false.groupby(['oa_category'])['upw_doi'].count() \\\n",
    "                    .plot.pie(figsize=(6,6), \n",
    "                              autopct='%1.0f%%',\n",
    "                              textprops={'fontsize': 11, 'fontweight': 400},\n",
    "                              colors=['#6da7de','#b3b3b3','#DFC86D'],\n",
    "                              shadow=True,\n",
    "                              startangle=90,\n",
    "                              explode=(0.06, 0, 0))\n",
    "\n",
    "ax.axis('equal')\n",
    "ax.set(ylabel='')\n",
    "\n",
    "plt.title('Proportion of wrong Classifications to Open Access Category',\n",
    "          fontdict={'fontsize': 12, 'fontweight': 700}, pad=20, x=0.5)\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "While a decrease in incorrect `other_repo` classifications can be observed, false recognition of `opendoar_other` has increased compared to the result from the previous analysis. In total, five repositories were classified incorrectly as `opendoar_other`. This time there were no faulty closed access cases, however two articles marked as `not_oa` were available as `free access` on the websites, see [Link](https://onlinelibrary.wiley.com/doi/abs/10.1002/humu.21197)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "ax = df_false.groupby(['revision'])['upw_doi'].count() \\\n",
    "                .plot.pie(figsize=(6,6), \n",
    "                          autopct='%1.0f%%',\n",
    "                          textprops={'fontsize': 11, 'fontweight': 400},\n",
    "                          colors=['#DEE000','#B5DF6D','#6da7de', '#6D79DF', '#DFC86D'],\n",
    "                          shadow=True,\n",
    "                          explode=(0, 0.1, 0, 0, 0),\n",
    "                          startangle=90)\n",
    "\n",
    "plt.title('Proportion of revised Classifications to Open Access Category',\n",
    "          fontdict={'fontsize': 12, 'fontweight': 700}, pad=30, x=0.5)\n",
    "\n",
    "ax.axis('equal')\n",
    "ax.set(ylabel='')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The figure shows that in over a third of the incorrect classifications a connection to institutional repositories has been proved. In one case a DOI led to a deleted document (noted as `none`). Though strictly speaking, the location was classified correctly. In two cases, no evidence to OpenDOAR could be proven. In both cases it was striking that the respective values in `url_for_landing_page` and `url_for_pdf` refer to different domains. The related repository was [CiteSeerX.psu](https://citeseerx.ist.psu.edu/index). Again classification problems arose with this [repository](https://v2.sherpa.ac.uk/id/repository/1874?template=opendoar). The repository was wrongly classified twice as `opendoar_subject` and once correctly as `opendoar_other`. Also the repository [DESY Publication Database](https://v2.sherpa.ac.uk/id/repository/1108), which was already incorrectly classified in the last analysis, was recognized as `other_repo`, but is found in OpenDOAR. The [Aberystwyth Research Portal](https://v2.sherpa.ac.uk/id/repository/4170) respository has been classified as `opendoar_other` in two cases, but is listed as an institutional repository in OpenDOAR. In one case, Unpaywall´s classification of a location was incorrect. Once again, a journal was incorrectly classified as a repository. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 504x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure(figsize=(7,4))\n",
    "\n",
    "df_false_edges = df_false.agg(lambda x: tuple([x['oa_category'], x['revision']]), \n",
    "                                      axis=1).values\n",
    "\n",
    "G = nx.DiGraph()\n",
    "G.add_edges_from(df_false_edges)\n",
    "\n",
    "for u, v, d in G.edges(data=True):\n",
    "    weight = list(df_false_edges).count(tuple([u, v]))\n",
    "    G[u][v]['weight'] = weight\n",
    "    \n",
    "edgewidth = [d['weight'] for (u, v, d) in G.edges(data=True)]\n",
    "edgewidth = [w / np.mean(edgewidth) for w in edgewidth]\n",
    "\n",
    "pos = nx.circular_layout(G)\n",
    "\n",
    "nx.draw(G, \n",
    "        pos=pos, \n",
    "        width=edgewidth, \n",
    "        with_labels=True, \n",
    "        node_size=900,\n",
    "        linewidths=0.3,\n",
    "        alpha=0.95,\n",
    "        node_color='#b3b3b3',\n",
    "        font_size=10, \n",
    "        font_weight='bold')\n",
    "\n",
    "ax = plt.gca() \n",
    "ax.collections[0].set_edgecolor('black')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Although the results should be viewed with caution, since two entries could possibly be listed in OpenDOAR and a repository that has been classified correctly but contains a deleted entry, an improved classification with respect to the sample was achieved. One can also see that there was no major discrepancy between incorrect classifications among OpenDOAR and non-OpenDOAR classifications."
   ]
  },
  {
   "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": "3.7.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}