Merge pull request #41 from PioneersHub/speaker_company_statistics

add speaker statistics

notebooks/pyconde-pydata-darmstadt-2025/10_submission_stats_v1.ipynb

@@ -100,6 +100,150 @@
     "spkrs_df = speakers_as_df(spkrs, with_questions=True)"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Speaker Statistics\n",
+    "\n",
+    "Determine the number of speakers per company."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Group spkrs_df by \"Q: Company / Institute\" and count the number of submission and return a dataframe\n",
+    "company_counts = spkrs_df.groupby('Q: Company / Institute').size().reset_index(name='count')\n",
+    "\n",
+    "# Sort the dataframe by the count of submissions\n",
+    "company_counts = company_counts.sort_values('count', ascending=False)\n",
+    "\n",
+    "# rename \"Q: Company / Institute\" to \"company_name\"\n",
+    "company_counts = company_counts.rename(columns={'Q: Company / Institute': 'company_name'})\n",
+    "\n",
+    "# company_counts.to_csv('company_counts.csv', index=False)\n",
+    "company_counts.head(10)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from difflib import get_close_matches\n",
+    "\n",
+    "# Step 0: Create a dictionary to map similar names to a consistent name\n",
+    "company_names = company_counts['company_name'].unique()\n",
+    "company_name_map = {}\n",
+    "for company_name in company_names:\n",
+    "    company_names_exclude_name = [name for name in company_names if name != company_name]\n",
+    "    \n",
+    "    # except handling company name\n",
+    "    lookup_name = company_name.strip()\n",
+    "\n",
+    "    # do lookup without company form\n",
+    "    if lookup_name.endswith('GmbH'):\n",
+    "        lookup_name = lookup_name[:-4]\n",
+    "    matches = get_close_matches(lookup_name, company_names_exclude_name, n=1, cutoff=.8)  # Adjust the cutoff as needed\n",
+    "    if matches:\n",
+    "        company_name_map[company_name] = matches[0]\n",
+    "\n",
+    "    # do lookup without company form\n",
+    "    if not lookup_name.endswith('GmbH'):\n",
+    "        lookup_name = f'{lookup_name} GmbH'\n",
+    "    matches = get_close_matches(lookup_name, company_names_exclude_name, n=1, cutoff=.8)  # Adjust the cutoff as needed\n",
+    "    if matches:\n",
+    "        company_name_map[company_name] = matches[0]\n",
+    "\n",
+    "    # do lookup with original company name\n",
+    "    matches = get_close_matches(company_name, company_names_exclude_name, n=1, cutoff=.8)  # Adjust the cutoff as needed\n",
+    "    if matches:\n",
+    "        company_name_map[company_name] = matches[0]\n",
+    "\n",
+    "company_name_map"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from collections import defaultdict\n",
+    "\n",
+    "# Step 1: Build an adjacency list to represent the graph\n",
+    "adjacency_list = defaultdict(set)\n",
+    "for key, value in company_name_map.items():\n",
+    "    adjacency_list[key].add(value)\n",
+    "    adjacency_list[value].add(key)\n",
+    "\n",
+    "# Step 2: Perform a DFS or BFS to find all connected components\n",
+    "def find_groups(adjacency_list):\n",
+    "    visited = set()\n",
+    "    groups = []\n",
+    "\n",
+    "    def dfs(node, group):\n",
+    "        visited.add(node)\n",
+    "        group.append(node)\n",
+    "        for neighbor in adjacency_list[node]:\n",
+    "            if neighbor not in visited:\n",
+    "                dfs(neighbor, group)\n",
+    "\n",
+    "    for node in adjacency_list:\n",
+    "        if node not in visited:\n",
+    "            group = []\n",
+    "            dfs(node, group)\n",
+    "            groups.append(group)\n",
+    "    \n",
+    "    return groups\n",
+    "\n",
+    "# Step 3: Get the grouped names\n",
+    "company_name_groups = find_groups(adjacency_list)\n",
+    "\n",
+    "# Step 4: for each key in company_name_map find its group in company_name_groups\n",
+    "company_name_group_map = {}\n",
+    "for key, value in company_name_map.items():\n",
+    "    for group in company_name_groups:\n",
+    "        if key in group:\n",
+    "            company_name_group_map[key] = group\n",
+    "            break\n",
+    "\n",
+    "# Step 5: Assign the group names if available\n",
+    "company_counts['company_name_grouping_arr'] = company_counts['company_name'].apply(lambda x: company_name_group_map.get(x, [x]))\n",
+    "\n",
+    "company_counts['company_name_grouping_str'] = company_counts['company_name_grouping_arr'].apply(lambda x: '[' + ', '.join(x) + ']')\n",
+    "\n",
+    "company_counts.head(10)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Group \"company_counts\" by company_name_grouping column and sum the count column\n",
+    "company_counts_by_group = company_counts.groupby('company_name_grouping_str')['count'].sum().reset_index()\n",
+    "\n",
+    "# Sort the dataframe by the count of submissions\n",
+    "company_counts_by_group = company_counts_by_group.sort_values('count', ascending=False)\n",
+    "\n",
+    "# company_counts_by_group.to_csv('company_counts_by_group.csv', index=False)\n",
+    "company_counts_by_group"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Submission Statistics\n",
+    "Determine number of speakers/submissions per track and main track."
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
@@ -289,7 +433,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.12.7"
+   "version": "3.10.16"
   }
  },
  "nbformat": 4,