{
"cells": [
{
"cell_type": "markdown",
"id": "d437d6f8-00a9-466a-9c8d-c1bcda6811a1",
"metadata": {},
"source": [
"# CX Machine Translation services usage analysis\n",
"\n",
"**Krishna Chaitanya Velaga, Data Scientist III, Wikimedia Foundation**\n",
"\n",
"**Last Update: 9 November 2023**\n",
"\n",
"[Ticket](https://phabricator.wikimedia.org/T338606) | [Codebase](https://github.com/wikimedia-research/machine-translation-service-analysis)"
]
},
{
"cell_type": "markdown",
"id": "7f8118de-fedf-4a89-b0c3-c4483a729201",
"metadata": {},
"source": [
"**Note** {content} \"\n",
" \n",
" centered_html = f\"{content}
\"\n",
" \n",
" display(HTML(centered_html))\n",
"\n",
"# format column as a percentage in a dataframe\n",
"def format_percent(column, df):\n",
" return df[column].map('{:.2%}'.format)\n",
"\n",
"# display dataframes horizontally with title for each\n",
"def display_h(frames, space=100):\n",
" html = \"\"\n",
" \n",
" for key in frames.keys():\n",
" html_df =f'{key} {frames[key]._repr_html_()}
'\n",
" html += html_df\n",
" \n",
" html = f\"\"\"\n",
" \n",
" {html}\n",
"
\"\"\"\n",
" \n",
" display_html(html, raw=True)"
]
},
{
"cell_type": "code",
"execution_count": 4,
"id": "06f88002-4530-437c-8998-4f5270cdcc01",
"metadata": {},
"outputs": [],
"source": [
"# connect to database\n",
"conn = duckdb.connect('data_gathering/secrets/mt_data.db')"
]
},
{
"cell_type": "code",
"execution_count": 5,
"id": "4119c009-7299-44a2-9e6e-3e0aff788d10",
"metadata": {},
"outputs": [],
"source": [
"def query(query_string, df=False, conn=conn):\n",
" if df==True:\n",
" return conn.sql(query_string).df()\n",
" else:\n",
" return conn.sql(query_string).show()"
]
},
{
"cell_type": "code",
"execution_count": 6,
"id": "1bacb4b0-ad34-4b0d-bf91-b5e94cc18647",
"metadata": {},
"outputs": [],
"source": [
"# conn.close()"
]
},
{
"cell_type": "markdown",
"id": "9c7f0d56-040b-4466-bac8-84f3e8423fa1",
"metadata": {},
"source": [
"# Overall-usage-across-all-languages"
]
},
{
"cell_type": "code",
"execution_count": 6,
"id": "4ca54a0d-2ffe-4545-a1bc-e8ddcdf64959",
"metadata": {},
"outputs": [],
"source": [
"# get 2023 overall data\n",
"mt_compare_overall_2023 = query(\"\"\"\n",
"WITH\n",
" base AS (\n",
" SELECT \n",
" mt_service,\n",
" COUNT(DISTINCT translation_id) AS n_translations\n",
" FROM \n",
" mt_logs\n",
" WHERE\n",
" translation_start_time >= '2023-08-01' AND\n",
" translation_start_time <= '2023-10-31'\n",
" GROUP BY\n",
" mt_service\n",
" ORDER BY\n",
" n_translations DESC)\n",
"\n",
"\n",
"SELECT\n",
" mt_service AS 'Machine translation service',\n",
" n_translations AS 'Number of Translations',\n",
" n_translations / SUM(n_translations) OVER () AS 'Percent of all published translations'\n",
"FROM\n",
" base\n",
"\"\"\", True)"
]
},
{
"cell_type": "code",
"execution_count": 7,
"id": "710750ca-0502-4f55-8637-ab2533bc2023",
"metadata": {},
"outputs": [],
"source": [
"mt_compare_overall_2023_pr = mt_compare_overall_2023.copy()\n",
"mt_compare_overall_2023_pr['Percent of all published translations'] = format_percent('Percent of all published translations', mt_compare_overall_2023_pr)"
]
},
{
"cell_type": "code",
"execution_count": 8,
"id": "42994d55-33ea-4df2-a133-0dc9640dd4ab",
"metadata": {
"tags": []
},
"outputs": [
{
"data": {
"text/html": [
"Published translations by machine translations service across all language pairs
"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"August to October 2023
"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"\n",
" \n",
"
\n",
"\n",
"
\n",
" \n",
" \n",
" Number of Translations \n",
" Percent of all published translations \n",
" \n",
" \n",
" Machine translation service \n",
" \n",
" \n",
" \n",
" \n",
" Google \n",
" 63335 \n",
" 80.28% \n",
" \n",
" \n",
" MinT \n",
" 6300 \n",
" 7.99% \n",
" \n",
" \n",
" scratch \n",
" 4704 \n",
" 5.96% \n",
" \n",
" \n",
" Apertium \n",
" 2144 \n",
" 2.72% \n",
" \n",
" \n",
" Yandex \n",
" 1505 \n",
" 1.91% \n",
" \n",
" \n",
" Elia \n",
" 865 \n",
" 1.10% \n",
" \n",
" \n",
" LingoCloud \n",
" 38 \n",
" 0.05% \n",
" \n",
" \n",
"
\n",
"
\n",
"
"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"# bar chart\n",
"mt_compare_overall_2023['Percent of all published translations'] *= 100\n",
"\n",
"plt.figure(figsize=(10, 6))\n",
"ax = sns.barplot(data=mt_compare_overall_2023, x='Machine translation service', y='Percent of all published translations', palette=\"BuGn_r\")\n",
"\n",
"# format y-axis as percetange\n",
"ax.yaxis.set_major_formatter(mtick.PercentFormatter(decimals=0))\n",
"\n",
"# add data labels for bars\n",
"for p in ax.patches:\n",
" ax.annotate(f'{p.get_height():.2f}%', (p.get_x() + p.get_width() / 2., p.get_height()), \n",
" ha='center', va='center', fontsize=8, fontweight='bold', color='black', \n",
" xytext=(0, 10), textcoords='offset points')\n",
"\n",
"# plot and axes titles\n",
"plt.title(\"Published translations by machine translations service across all language pairs\\nfrom August to October 2023\")\n",
"plt.xlabel(\"Machine translation service\", fontweight='bold')\n",
"plt.ylabel(\"Percent of all published translations\", fontweight='bold')\n",
"\n",
"plt.show()"
]
},
{
"cell_type": "markdown",
"id": "f37a9366-4ea9-4e68-8c5c-47d61933c9ff",
"metadata": {},
"source": [
"## YoY-comparision"
]
},
{
"cell_type": "code",
"execution_count": 10,
"id": "9060983c-c2d1-4513-9015-e44d9250c8c6",
"metadata": {},
"outputs": [],
"source": [
"# get 2022 overall data\n",
"mt_compare_overall_2022 = query(\"\"\"\n",
"WITH\n",
" base AS (\n",
" SELECT \n",
" mt_service,\n",
" COUNT(DISTINCT translation_id) AS n_translations\n",
" FROM \n",
" mt_logs\n",
" WHERE\n",
" translation_start_time >= '2022-08-01' AND\n",
" translation_start_time <= '2022-10-31'\n",
" GROUP BY\n",
" mt_service\n",
" ORDER BY\n",
" n_translations DESC)\n",
"\n",
"\n",
"SELECT\n",
" mt_service AS 'Machine translation service',\n",
" n_translations AS 'Number of Translations',\n",
" n_translations / SUM(n_translations) OVER () AS 'Percent of all published translations'\n",
"FROM\n",
" base\n",
"\"\"\", True)"
]
},
{
"cell_type": "code",
"execution_count": 11,
"id": "c91a4599-1e0f-44a2-bce8-75cbe16ba7a0",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"Published translations by machine translations service across all language pairs
"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"\n",
" \n",
"
August to October 2022
\n",
"\n",
"
\n",
" \n",
" \n",
" Number of Translations \n",
" Percent of all published translations \n",
" \n",
" \n",
" Machine translation service \n",
" \n",
" \n",
" \n",
" \n",
" Google \n",
" 69824 \n",
" 82.72% \n",
" \n",
" \n",
" NLLB-200 \n",
" 5534 \n",
" 6.56% \n",
" \n",
" \n",
" scratch \n",
" 3270 \n",
" 3.87% \n",
" \n",
" \n",
" Yandex \n",
" 2965 \n",
" 3.51% \n",
" \n",
" \n",
" Apertium \n",
" 1576 \n",
" 1.87% \n",
" \n",
" \n",
" Elia \n",
" 978 \n",
" 1.16% \n",
" \n",
" \n",
" OpusMT \n",
" 214 \n",
" 0.25% \n",
" \n",
" \n",
" LingoCloud \n",
" 25 \n",
" 0.03% \n",
" \n",
" \n",
" MinT \n",
" 19 \n",
" 0.02% \n",
" \n",
" \n",
"
\n",
"
August to October 2023
\n",
"\n",
"
\n",
" \n",
" \n",
" Number of Translations \n",
" Percent of all published translations \n",
" \n",
" \n",
" Machine translation service \n",
" \n",
" \n",
" \n",
" \n",
" Google \n",
" 63335 \n",
" 80.28% \n",
" \n",
" \n",
" MinT \n",
" 6300 \n",
" 7.99% \n",
" \n",
" \n",
" scratch \n",
" 4704 \n",
" 5.96% \n",
" \n",
" \n",
" Apertium \n",
" 2144 \n",
" 2.72% \n",
" \n",
" \n",
" Yandex \n",
" 1505 \n",
" 1.91% \n",
" \n",
" \n",
" Elia \n",
" 865 \n",
" 1.10% \n",
" \n",
" \n",
" LingoCloud \n",
" 38 \n",
" 0.05% \n",
" \n",
" \n",
"
\n",
"
\n",
"
Number of Daily Translations (Median)
"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"\n",
" \n",
"
August to October 2022
\n",
"\n",
"
\n",
" \n",
" \n",
" # Translations \n",
" \n",
" \n",
" Service \n",
" \n",
" \n",
" \n",
" \n",
" Google \n",
" 719 \n",
" \n",
" \n",
" scratch \n",
" 33 \n",
" \n",
" \n",
" Yandex \n",
" 27 \n",
" \n",
" \n",
" NLLB-200 \n",
" 20 \n",
" \n",
" \n",
" Apertium \n",
" 16 \n",
" \n",
" \n",
" Elia \n",
" 8 \n",
" \n",
" \n",
" OpusMT \n",
" 3 \n",
" \n",
" \n",
" LingoCloud \n",
" 1 \n",
" \n",
" \n",
" MinT \n",
" 1 \n",
" \n",
" \n",
"
\n",
"
April to June 2023
\n",
"\n",
"
\n",
" \n",
" \n",
" # Translations \n",
" \n",
" \n",
" Service \n",
" \n",
" \n",
" \n",
" \n",
" Google \n",
" 708 \n",
" \n",
" \n",
" scratch \n",
" 51 \n",
" \n",
" \n",
" Yandex \n",
" 26 \n",
" \n",
" \n",
" NLLB-200 \n",
" 24 \n",
" \n",
" \n",
" Apertium \n",
" 20 \n",
" \n",
" \n",
" MinT \n",
" 19 \n",
" \n",
" \n",
" Elia \n",
" 9 \n",
" \n",
" \n",
" OpusMT \n",
" 2 \n",
" \n",
" \n",
" LingoCloud \n",
" 1 \n",
" \n",
" \n",
"
\n",
"
August to October 2023
\n",
"\n",
"
\n",
" \n",
" \n",
" # Translations \n",
" \n",
" \n",
" Service \n",
" \n",
" \n",
" \n",
" \n",
" Google \n",
" 669 \n",
" \n",
" \n",
" scratch \n",
" 49 \n",
" \n",
" \n",
" MinT \n",
" 48 \n",
" \n",
" \n",
" Apertium \n",
" 23 \n",
" \n",
" \n",
" Yandex \n",
" 15 \n",
" \n",
" \n",
" Elia \n",
" 8 \n",
" \n",
" \n",
" LingoCloud \n",
" 1 \n",
" \n",
" \n",
"
\n",
"
\n",
"
Languge pairs where an optional service was used more or close to the default
"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"(having at least five published translations)
"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"\n",
" \n",
"
\n",
"\n",
"
\n",
" \n",
" \n",
" Target \n",
" # Translations \n",
" % of Translations \n",
" pair \n",
" MT service used \n",
" \n",
" \n",
" Default service \n",
" Source \n",
" \n",
" \n",
" \n",
" \n",
" MinT \n",
" en \n",
" zu \n",
" 55 \n",
" 96.49% \n",
" en-zu \n",
" Google \n",
" \n",
" \n",
" Google \n",
" en \n",
" ja \n",
" 57 \n",
" 100.00% \n",
" en-ja \n",
" scratch \n",
" \n",
" \n",
" tr \n",
" en \n",
" 7 \n",
" 100.00% \n",
" tr-en \n",
" scratch \n",
" \n",
" \n",
" en \n",
" de \n",
" 16 \n",
" 100.00% \n",
" en-de \n",
" scratch \n",
" \n",
" \n",
" ar \n",
" en \n",
" 31 \n",
" 93.94% \n",
" ar-en \n",
" scratch \n",
" \n",
" \n",
" de \n",
" lb \n",
" 29 \n",
" 90.62% \n",
" de-lb \n",
" Yandex \n",
" \n",
" \n",
" fa \n",
" en \n",
" 45 \n",
" 75.00% \n",
" fa-en \n",
" scratch \n",
" \n",
" \n",
" he \n",
" en \n",
" 27 \n",
" 75.00% \n",
" he-en \n",
" scratch \n",
" \n",
" \n",
" uk \n",
" be \n",
" 52 \n",
" 72.22% \n",
" uk-be \n",
" Yandex \n",
" \n",
" \n",
" en \n",
" st \n",
" 9 \n",
" 60.00% \n",
" en-st \n",
" MinT \n",
" \n",
" \n",
" he \n",
" yi \n",
" 9 \n",
" 52.94% \n",
" he-yi \n",
" Yandex \n",
" \n",
" \n",
" fr \n",
" en \n",
" 13 \n",
" 50.00% \n",
" fr-en \n",
" scratch \n",
" \n",
" \n",
" be \n",
" uk \n",
" 11 \n",
" 50.00% \n",
" be-uk \n",
" Yandex \n",
" \n",
" \n",
" en \n",
" id \n",
" 596 \n",
" 44.81% \n",
" en-id \n",
" scratch \n",
" \n",
" \n",
" en \n",
" ku \n",
" 9 \n",
" 37.50% \n",
" en-ku \n",
" MinT \n",
" \n",
" \n",
" ms \n",
" simple \n",
" 7 \n",
" 33.33% \n",
" ms-simple \n",
" MinT \n",
" \n",
" \n",
" ar \n",
" om \n",
" 9 \n",
" 29.03% \n",
" ar-om \n",
" scratch \n",
" \n",
" \n",
" en \n",
" fa \n",
" 705 \n",
" 27.91% \n",
" en-fa \n",
" scratch \n",
" \n",
" \n",
" it \n",
" fr \n",
" 12 \n",
" 25.53% \n",
" it-fr \n",
" scratch \n",
" \n",
" \n",
" Elia \n",
" es \n",
" eu \n",
" 241 \n",
" 31.67% \n",
" es-eu \n",
" Google \n",
" \n",
" \n",
" en \n",
" eu \n",
" 70 \n",
" 28.46% \n",
" en-eu \n",
" Google \n",
" \n",
" \n",
" Apertium \n",
" es \n",
" en \n",
" 19 \n",
" 67.86% \n",
" es-en \n",
" scratch \n",
" \n",
" \n",
" nl \n",
" af \n",
" 9 \n",
" 50.00% \n",
" nl-af \n",
" Google \n",
" \n",
" \n",
" be \n",
" ru \n",
" 10 \n",
" 35.71% \n",
" be-ru \n",
" Google \n",
" \n",
" \n",
" en \n",
" gl \n",
" 81 \n",
" 33.06% \n",
" en-gl \n",
" Google \n",
" \n",
" \n",
" es \n",
" ca \n",
" 169 \n",
" 32.82% \n",
" es-ca \n",
" Google \n",
" \n",
" \n",
"
\n",
"
\n",
"
Percentage of articles deleted, by Machine Translation Service
"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"August to October 2023
"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"\n",
" \n",
"
\n",
"\n",
"
\n",
" \n",
" \n",
" Number of Articles Deleted \n",
" Number of Articles Created \n",
" Percentage of deleted articles \n",
" Percent of translations modified under 10% \n",
" \n",
" \n",
" Machine Translation Service \n",
" \n",
" \n",
" \n",
" \n",
" Google \n",
" 2926 \n",
" 56201 \n",
" 5.21% \n",
" 30.34% \n",
" \n",
" \n",
" MinT \n",
" 97 \n",
" 5655 \n",
" 1.72% \n",
" 36.86% \n",
" \n",
" \n",
" scratch \n",
" 166 \n",
" 4185 \n",
" 3.97% \n",
" n/a \n",
" \n",
" \n",
" Apertium \n",
" 56 \n",
" 1994 \n",
" 2.81% \n",
" 22.81% \n",
" \n",
" \n",
" Yandex \n",
" 81 \n",
" 1299 \n",
" 6.24% \n",
" 29.17% \n",
" \n",
" \n",
" Elia \n",
" 15 \n",
" 811 \n",
" 1.85% \n",
" 17.92% \n",
" \n",
" \n",
" LingoCloud \n",
" 0 \n",
" 35 \n",
" 0.00% \n",
" 21.05% \n",
" \n",
" \n",
"
\n",
"
\n",
"
Languages pairs where has been a major change in percentage of translations made using Google
"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"with minimum 25 published translations and the difference is more than 40%
"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"\n",
" \n",
"
\n",
"\n",
"
\n",
" \n",
" \n",
" mt_service \n",
" # Translations 2022 \n",
" # Translations 2023 \n",
" % of Translations 2022 \n",
" % of Translations 2023 \n",
" % Difference of all Translations \n",
" \n",
" \n",
" Source \n",
" Target \n",
" \n",
" \n",
" \n",
" \n",
" ar \n",
" ckb \n",
" Google \n",
" 41.0 \n",
" 237.0 \n",
" 14.75% \n",
" 85.25% \n",
" 70.50% \n",
" \n",
" \n",
" az \n",
" ka \n",
" Google \n",
" 46.0 \n",
" 4.0 \n",
" 92.00% \n",
" 8.00% \n",
" -84.00% \n",
" \n",
" \n",
" be \n",
" ru \n",
" Google \n",
" 321.0 \n",
" 10.0 \n",
" 96.98% \n",
" 3.02% \n",
" -93.96% \n",
" \n",
" \n",
" bs \n",
" sr \n",
" Google \n",
" 83.0 \n",
" 739.0 \n",
" 10.10% \n",
" 89.90% \n",
" 79.81% \n",
" \n",
" \n",
" ceb \n",
" es \n",
" Google \n",
" 73.0 \n",
" 19.0 \n",
" 79.35% \n",
" 20.65% \n",
" -58.70% \n",
" \n",
" \n",
" de \n",
" fa \n",
" Google \n",
" 49.0 \n",
" 1.0 \n",
" 98.00% \n",
" 2.00% \n",
" -96.00% \n",
" \n",
" \n",
" hr \n",
" Google \n",
" 34.0 \n",
" 2.0 \n",
" 94.44% \n",
" 5.56% \n",
" -88.89% \n",
" \n",
" \n",
" it \n",
" Google \n",
" 77.0 \n",
" 32.0 \n",
" 70.64% \n",
" 29.36% \n",
" -41.28% \n",
" \n",
" \n",
" ku \n",
" Google \n",
" 36.0 \n",
" 1.0 \n",
" 97.30% \n",
" 2.70% \n",
" -94.59% \n",
" \n",
" \n",
" nl \n",
" Google \n",
" 59.0 \n",
" 17.0 \n",
" 77.63% \n",
" 22.37% \n",
" -55.26% \n",
" \n",
" \n",
" ru \n",
" Google \n",
" 44.0 \n",
" 14.0 \n",
" 75.86% \n",
" 24.14% \n",
" -51.72% \n",
" \n",
" \n",
" tr \n",
" Google \n",
" 42.0 \n",
" 14.0 \n",
" 75.00% \n",
" 25.00% \n",
" -50.00% \n",
" \n",
" \n",
" en \n",
" af \n",
" Google \n",
" 79.0 \n",
" 222.0 \n",
" 26.25% \n",
" 73.75% \n",
" 47.51% \n",
" \n",
" \n",
" be \n",
" Google \n",
" 27.0 \n",
" 146.0 \n",
" 15.61% \n",
" 84.39% \n",
" 68.79% \n",
" \n",
" \n",
" da \n",
" Google \n",
" 119.0 \n",
" 337.0 \n",
" 26.10% \n",
" 73.90% \n",
" 47.81% \n",
" \n",
" \n",
" fi \n",
" Google \n",
" 48.0 \n",
" 8.0 \n",
" 85.71% \n",
" 14.29% \n",
" -71.43% \n",
" \n",
" \n",
" ha \n",
" Google \n",
" 1230.0 \n",
" 3528.0 \n",
" 25.85% \n",
" 74.15% \n",
" 48.30% \n",
" \n",
" \n",
" id \n",
" Google \n",
" 156.0 \n",
" 725.0 \n",
" 17.71% \n",
" 82.29% \n",
" 64.59% \n",
" \n",
" \n",
" ig \n",
" Google \n",
" 28.0 \n",
" 743.0 \n",
" 3.63% \n",
" 96.37% \n",
" 92.74% \n",
" \n",
" \n",
" ku \n",
" Google \n",
" 148.0 \n",
" 15.0 \n",
" 90.80% \n",
" 9.20% \n",
" -81.60% \n",
" \n",
" \n",
" sk \n",
" Google \n",
" 30.0 \n",
" 103.0 \n",
" 22.56% \n",
" 77.44% \n",
" 54.89% \n",
" \n",
" \n",
" sq \n",
" Google \n",
" 200.0 \n",
" 727.0 \n",
" 21.57% \n",
" 78.43% \n",
" 56.85% \n",
" \n",
" \n",
" sw \n",
" Google \n",
" 34.0 \n",
" 4.0 \n",
" 89.47% \n",
" 10.53% \n",
" -78.95% \n",
" \n",
" \n",
" te \n",
" Google \n",
" 316.0 \n",
" 1867.0 \n",
" 14.48% \n",
" 85.52% \n",
" 71.05% \n",
" \n",
" \n",
" tg \n",
" Google \n",
" 31.0 \n",
" 1.0 \n",
" 96.88% \n",
" 3.12% \n",
" -93.75% \n",
" \n",
" \n",
" th \n",
" Google \n",
" 26.0 \n",
" 230.0 \n",
" 10.16% \n",
" 89.84% \n",
" 79.69% \n",
" \n",
" \n",
" uz \n",
" Google \n",
" 13108.0 \n",
" 3132.0 \n",
" 80.71% \n",
" 19.29% \n",
" -61.43% \n",
" \n",
" \n",
" es \n",
" ar \n",
" Google \n",
" 36.0 \n",
" 14.0 \n",
" 72.00% \n",
" 28.00% \n",
" -44.00% \n",
" \n",
" \n",
" bn \n",
" Google \n",
" 108.0 \n",
" 6.0 \n",
" 94.74% \n",
" 5.26% \n",
" -89.47% \n",
" \n",
" \n",
" fa \n",
" Google \n",
" 59.0 \n",
" 1.0 \n",
" 98.33% \n",
" 1.67% \n",
" -96.67% \n",
" \n",
" \n",
" fr \n",
" ar \n",
" Google \n",
" 63.0 \n",
" 346.0 \n",
" 15.40% \n",
" 84.60% \n",
" 69.19% \n",
" \n",
" \n",
" cs \n",
" Google \n",
" 62.0 \n",
" 8.0 \n",
" 88.57% \n",
" 11.43% \n",
" -77.14% \n",
" \n",
" \n",
" gl \n",
" es \n",
" Google \n",
" 106.0 \n",
" 30.0 \n",
" 77.94% \n",
" 22.06% \n",
" -55.88% \n",
" \n",
" \n",
" he \n",
" eo \n",
" Google \n",
" 67.0 \n",
" 2.0 \n",
" 97.10% \n",
" 2.90% \n",
" -94.20% \n",
" \n",
" \n",
" hu \n",
" he \n",
" Google \n",
" 61.0 \n",
" 209.0 \n",
" 22.59% \n",
" 77.41% \n",
" 54.81% \n",
" \n",
" \n",
" it \n",
" fr \n",
" Google \n",
" 204.0 \n",
" 35.0 \n",
" 85.36% \n",
" 14.64% \n",
" -70.71% \n",
" \n",
" \n",
" ja \n",
" zh \n",
" Google \n",
" 84.0 \n",
" 228.0 \n",
" 26.92% \n",
" 73.08% \n",
" 46.15% \n",
" \n",
" \n",
" kk \n",
" uz \n",
" Google \n",
" 241.0 \n",
" 4.0 \n",
" 98.37% \n",
" 1.63% \n",
" -96.73% \n",
" \n",
" \n",
" pl \n",
" uk \n",
" Google \n",
" 51.0 \n",
" 239.0 \n",
" 17.59% \n",
" 82.41% \n",
" 64.83% \n",
" \n",
" \n",
" ru \n",
" el \n",
" Google \n",
" 31.0 \n",
" 1.0 \n",
" 96.88% \n",
" 3.12% \n",
" -93.75% \n",
" \n",
" \n",
" fr \n",
" Google \n",
" 27.0 \n",
" 73.0 \n",
" 27.00% \n",
" 73.00% \n",
" 46.00% \n",
" \n",
" \n",
" ka \n",
" Google \n",
" 28.0 \n",
" 7.0 \n",
" 80.00% \n",
" 20.00% \n",
" -60.00% \n",
" \n",
" \n",
" sr \n",
" Google \n",
" 30.0 \n",
" 3.0 \n",
" 90.91% \n",
" 9.09% \n",
" -81.82% \n",
" \n",
" \n",
" tg \n",
" Google \n",
" 69.0 \n",
" 18.0 \n",
" 79.31% \n",
" 20.69% \n",
" -58.62% \n",
" \n",
" \n",
" uz \n",
" Google \n",
" 4074.0 \n",
" 1211.0 \n",
" 77.09% \n",
" 22.91% \n",
" -54.17% \n",
" \n",
" \n",
" simple \n",
" uz \n",
" Google \n",
" 33.0 \n",
" 4.0 \n",
" 89.19% \n",
" 10.81% \n",
" -78.38% \n",
" \n",
" \n",
" tr \n",
" az \n",
" Google \n",
" 44.0 \n",
" 184.0 \n",
" 19.30% \n",
" 80.70% \n",
" 61.40% \n",
" \n",
" \n",
" uz \n",
" Google \n",
" 242.0 \n",
" 69.0 \n",
" 77.81% \n",
" 22.19% \n",
" -55.63% \n",
" \n",
" \n",
" uk \n",
" be \n",
" Google \n",
" 31.0 \n",
" 12.0 \n",
" 72.09% \n",
" 27.91% \n",
" -44.19% \n",
" \n",
" \n",
" ru \n",
" Google \n",
" 170.0 \n",
" 64.0 \n",
" 72.65% \n",
" 27.35% \n",
" -45.30% \n",
" \n",
" \n",
" ur \n",
" ps \n",
" Google \n",
" 263.0 \n",
" 31.0 \n",
" 89.46% \n",
" 10.54% \n",
" -78.91% \n",
" \n",
" \n",
" zh \n",
" es \n",
" Google \n",
" 27.0 \n",
" 3.0 \n",
" 90.00% \n",
" 10.00% \n",
" -80.00% \n",
" \n",
" \n",
"
\n",
"
\n",
"
Languages pairs where has been a major change in percentage of translations made using Yandex
"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"with minimum 25 published translations and the difference is more than 25%
"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"\n",
" \n",
"
\n",
"\n",
"
\n",
" \n",
" \n",
" # Translations 2022 \n",
" # Translations 2023 \n",
" % of Translations 2022 \n",
" % of Translations 2023 \n",
" % Difference of all Translations \n",
" \n",
" \n",
" Source \n",
" Target \n",
" \n",
" \n",
" \n",
" \n",
" en \n",
" ar \n",
" 39.0 \n",
" 195.0 \n",
" 16.67% \n",
" 83.33% \n",
" 66.67% \n",
" \n",
" \n",
" bn \n",
" 107.0 \n",
" 43.0 \n",
" 71.33% \n",
" 28.67% \n",
" -42.67% \n",
" \n",
" \n",
" fa \n",
" 51.0 \n",
" 13.0 \n",
" 79.69% \n",
" 20.31% \n",
" -59.38% \n",
" \n",
" \n",
" he \n",
" 50.0 \n",
" 7.0 \n",
" 87.72% \n",
" 12.28% \n",
" -75.44% \n",
" \n",
" \n",
" nb \n",
" 182.0 \n",
" 2.0 \n",
" 98.91% \n",
" 1.09% \n",
" -97.83% \n",
" \n",
" \n",
" pa \n",
" 28.0 \n",
" 1.0 \n",
" 96.55% \n",
" 3.45% \n",
" -93.10% \n",
" \n",
" \n",
" tr \n",
" 42.0 \n",
" 21.0 \n",
" 66.67% \n",
" 33.33% \n",
" -33.33% \n",
" \n",
" \n",
" uz \n",
" 740.0 \n",
" 73.0 \n",
" 91.02% \n",
" 8.98% \n",
" -82.04% \n",
" \n",
" \n",
" ru \n",
" ba \n",
" 277.0 \n",
" 69.0 \n",
" 80.06% \n",
" 19.94% \n",
" -60.12% \n",
" \n",
" \n",
" be \n",
" 130.0 \n",
" 229.0 \n",
" 36.21% \n",
" 63.79% \n",
" 27.58% \n",
" \n",
" \n",
" uz \n",
" 152.0 \n",
" 33.0 \n",
" 82.16% \n",
" 17.84% \n",
" -64.32% \n",
" \n",
" \n",
" uk \n",
" be \n",
" 521.0 \n",
" 52.0 \n",
" 90.92% \n",
" 9.08% \n",
" -81.85% \n",
" \n",
" \n",
"
\n",
"
\n",
"
Languages pairs where has been a major change in percentage of translations made using Elia
"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"with minimum 25 published translations and the difference is more than 25%
"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"\n",
" \n",
"
\n",
"\n",
"
\n",
" \n",
" \n",
" # Translations 2022 \n",
" # Translations 2023 \n",
" % of Translations 2022 \n",
" % of Translations 2023 \n",
" % Difference of all Translations \n",
" \n",
" \n",
" Source \n",
" Target \n",
" \n",
" \n",
" \n",
" \n",
"
\n",
"
\n",
"
Languages pairs where has been a major change in percentage of translations made using Apertium
"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"with minimum 25 published translations and the difference is more than 25%
"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"\n",
" \n",
"
\n",
"\n",
"
\n",
" \n",
" \n",
" # Translations 2022 \n",
" # Translations 2023 \n",
" % of Translations 2022 \n",
" % of Translations 2023 \n",
" % Difference of all Translations \n",
" \n",
" \n",
" Source \n",
" Target \n",
" \n",
" \n",
" \n",
" \n",
" en \n",
" eo \n",
" 120.0 \n",
" 481.0 \n",
" 19.97% \n",
" 80.03% \n",
" 60.07% \n",
" \n",
" \n",
" es \n",
" gl \n",
" 96.0 \n",
" 290.0 \n",
" 24.87% \n",
" 75.13% \n",
" 50.26% \n",
" \n",
" \n",
"
\n",
"
\n",
"
Languages pairs where has been a major change in percentage of translations made without use of any machine translation
"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"with minimum 25 published translations and the difference is more than 25%
"
],
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"\n",
" \n",
"
\n",
"\n",
"
\n",
" \n",
" \n",
" # Translations 2022 \n",
" # Translations 2023 \n",
" % of Translations 2022 \n",
" % of Translations 2023 \n",
" % Difference of all Translations \n",
" \n",
" \n",
" Source \n",
" Target \n",
" \n",
" \n",
" \n",
" \n",
" en \n",
" ar \n",
" 193.0 \n",
" 500.0 \n",
" 27.85% \n",
" 72.15% \n",
" 44.30% \n",
" \n",
" \n",
" bn \n",
" 20.0 \n",
" 50.0 \n",
" 28.57% \n",
" 71.43% \n",
" 42.86% \n",
" \n",
" \n",
" el \n",
" 14.0 \n",
" 5.0 \n",
" 73.68% \n",
" 26.32% \n",
" -47.37% \n",
" \n",
" \n",
" id \n",
" 154.0 \n",
" 596.0 \n",
" 20.53% \n",
" 79.47% \n",
" 58.93% \n",
" \n",
" \n",
" ml \n",
" 28.0 \n",
" 4.0 \n",
" 87.50% \n",
" 12.50% \n",
" -75.00% \n",
" \n",
" \n",
" nb \n",
" 12.0 \n",
" 6.0 \n",
" 66.67% \n",
" 33.33% \n",
" -33.33% \n",
" \n",
" \n",
" pt \n",
" 58.0 \n",
" 184.0 \n",
" 23.97% \n",
" 76.03% \n",
" 52.07% \n",
" \n",
" \n",
" sv \n",
" 74.0 \n",
" 1.0 \n",
" 98.67% \n",
" 1.33% \n",
" -97.33% \n",
" \n",
" \n",
" uz \n",
" 241.0 \n",
" 28.0 \n",
" 89.59% \n",
" 10.41% \n",
" -79.18% \n",
" \n",
" \n",
" zh \n",
" 118.0 \n",
" 60.0 \n",
" 66.29% \n",
" 33.71% \n",
" -32.58% \n",
" \n",
" \n",
" es \n",
" fr \n",
" 25.0 \n",
" 12.0 \n",
" 67.57% \n",
" 32.43% \n",
" -35.14% \n",
" \n",
" \n",
" ru \n",
" az \n",
" 26.0 \n",
" 8.0 \n",
" 76.47% \n",
" 23.53% \n",
" -52.94% \n",
" \n",
" \n",
" be \n",
" 37.0 \n",
" 6.0 \n",
" 86.05% \n",
" 13.95% \n",
" -72.09% \n",
" \n",
" \n",
" uz \n",
" 11.0 \n",
" 68.0 \n",
" 13.92% \n",
" 86.08% \n",
" 72.15% \n",
" \n",
" \n",
" ur \n",
" en \n",
" 12.0 \n",
" 4.0 \n",
" 75.00% \n",
" 25.00% \n",
" -50.00% \n",
" \n",
" \n",
" zh \n",
" uk \n",
" 13.0 \n",
" 3.0 \n",
" 81.25% \n",
" 18.75% \n",
" -62.50% \n",
" \n",
" \n",
"
\n",
"
\n",
"