User:AKhatun/Wikidata Subgraph Query Analysis

Analysis on Subgraphs in Wikidata showed how large each of the subgraphs are in Wikidata and how connected they are. This page shows the results from analysis on the queries that relate to these subgraph. The questions that needed to be answered were:

• How many(percent) queries access each subgraph?
• How many queries access multiple subgraphs at once? i.e, how much overlap can we expect in subgraphs?
• How long do these queries take?
• How many user-agents access each subgraph? How many of them access lots of subgraphs, or are they confined to a small set of subgraphs? Do some of them dominate queries in multiple subgraphs?
• Are there chunks of similar queries in these subgraphs? i.e, how diverse the queries in each subgraph are.

Shorter TL;DR

• The monthly queries that touch the top 341 subgraphs were analyzed. The percentage of queries for each of the subgraphs change slightly by month. Only two subgraphs have significantly more queries compared to other subgraphs. Human subgraph has 30% of the queries, and Taxon subgraph consists of ~20% of the queries.
• Query times are divided into 5 classes. Most query times are low (10ms to 100ms, the second class of query time), some are 100ms to 1s (the third class of query time).
• Most subgraphs don’t have a lot of user-agents accessing them. Some of them have a few user-agents doing most of the queries. The trend of a dominating single source of queries is not widespread among subgraphs.
• Most user agents (89% of them) query one subgraph only. Few user agents query a lot of subgraphs as well.
• 70% of all queries touch on 1 or 2 subgraphs. 64% of all queries touch on only 1 subgraph.
• No significant correlation was found between query time and the number of subgraphs a query tries to access.

• For human subgraph:
  • The queries that were estimated to be related to the human subgraph accounted for 31.94% of all queries in Wikidata. 25.78% of queries used only the human subgraph and the rest 6.16% of queries used a mix of human and various other subgraphs.
  • The total query time of the human subgraph is 34% of the total query time. The average time per query is 0.3 seconds (300 ms). Most queries in this subgraph are small and simple.
  • More than 50% of the queries in human subgraph are done by ~10 user agents. So only a small number of user-agents do most of the queries.
  • The top 10 types of queries account for 60% of the queries of human subgraph. So a small type of queries cover the bulk of the queries.
  • Some user agents do a moderate amount of small simple queries of various types, but most user agents do only 1 type of query.

• For taxon subgraph:
  • The queries that were estimated to be related to the taxon subgraph accounted for 14.26% of all queries in Wikidata. 13.57% queries used only the taxon subgraph and the rest 0.69% queries used a mix of taxon and various other subgraphs
  • Q-ID (Q16521) match is almost the only reason for queries to be in the taxon subgraph.
  • Most properties that match the taxon subgraph are some sort of external IDs.
  • The total query time of taxon subgraph is ~3% of total query time. The average time per query is 0.064 seconds (64 ms). Almost all queries in this subgraph are small and simple.
  • The top user agent performs 85% of taxon subgraph queries. Basically, a single user agent does most of the queries.
  • The variety of query types in taxon subgraph is quite less (1.1K) compared to 11K query types in human subgraph. Only the top 3 types of queries account for ~90% of the queries of taxon subgraph.
  • Most user agents make only 1 type of query.
  • In sum, most queries in taxon subgraph are small and simple, take between 10 to 100ms time to run, and are mostly by 1 user agents.

TL;DR

• The monthly queries that touch/access each of the top 341 subgraphs was determined. The percentage of queries for each of the subgraphs change slightly by month. The list for the top 50 subgraphs for Nov, 2021 can be found here: #Query count and time, and comparison with Oct and Nov, 2021 data is shown here: #Comparison of subgraph queries across time.
• Only two subgraphs have significantly more queries compared to other subgraphs. Human subgraph has 30% of the queries, and Taxon subgraph consists of ~20% of the queries.
• Most subgraphs have most query times in the range of 10ms to 100ms. The second most common class is 100ms to 1s. A small number of subgraphs have more time-consuming queries (1s - 10s)
• Most subgraphs don’t have a lot of user-agents accessing them. Some of them have a few user-agents doing most of the queries. ~30 out of 341 subgraphs have a user agent that queries >=50% of all queries of that particular subgraphs. 6 subgraphs have a user agent doing 80-90% of the queries. So the trend of a dominating single source of queries is not widespread among subgraphs but is present in a few subgraphs nonetheless.
• Most user agents (89% of them) query one subgraph only. Some user agents query a lot of subgraphs as well. Explicit observations about some user-agents are shown in #User agent vs Subgraph section.
• Looking at the connection among subgraphs through queries we see: 70% of all queries (97% of queries in 341 subgraphs) touch on 1 or 2 subgraphs. 64% of all queries (90% of queries in 341 subgraphs) touch on only 1 subgraph.
• No significant correlation was found between query time and the number of subgraphs a query tries to access. But queries that access more subgraphs (although less in number) do appear slightly more in larger query time classes.
• In-depth analysis was done on Human and Taxon subgraph queries since they account for the most queries per subgraph.

• For human subgraph:
  • The queries that were estimated to be related to the human subgraph accounted for 31.94% of all queries in Wikidata. 25.78% of queries used only the human subgraph and the rest 6.16% of queries used a mix of human and various other subgraphs.
  • A lot of the queries are associated with human subgraph due to the properties they use, the instance of items, and URIs in subject or object.
  • There are some high-usage items, but mostly a long list of low-usage items. For properties, only ~10 properties account for most of the queries in the human subgraph. The matched URIs form a smooth logarithmic pattern, with most URIs being Wikipedia article links in various languages.
  • The total query time of the human subgraph is 34% of the total query time. The average time per query is 0.3 seconds (300 ms). Most queries in this subgraph are small and simple.
  • More than 50% of the queries in human subgraph are done by ~10 user agents. The top 2 user agents perform 20% of the human subgraph queries (7% of all queries). The first user agent does small queries that don’t require much time (10-100ms), whereas the second user agent performs queries that take comparatively more time (100ms - 1s). Note that user-agent strings are not directly representative of distinct users. The second top user-agent here, for example, has multiple variations of user-agent strings which were considered different at this time for simplicity.
  • In terms of query type (type is determined by the operations used in a query): the top 10 types of queries account for 60% of the queries of human subgraph. The rest form a long tail of less-used query types.
  • Most user agents do only 1 type of query. 8 user agents perform more than 500 types of queries. But looking into these types, it seems most of these types don’t have a lot of queries, and are mostly small simple queries. In short: Some user agents do a moderate amount of small simple queries of various types, and most user agents do only 1 type of query.
• For taxon subgraph:
  • The queries that were estimated to be related to the taxon subgraph accounted for 14.26% of all queries in Wikidata. 13.57% queries used only the taxon subgraph and the rest 0.69% queries used a mix of taxon and various other subgraphs
  • Q-ID match is almost the only reason for queries to be in the taxon subgraph (12% out of 14%). Basically, these are the queries that have Q16521 in them.
  • Only a few (~3) items match with significant number of queries, forming a logarithmic distribution of item use across queries. The distribution of properties used in taxon subgraph queries is also extremely skewed by only ~10 properties, with most of these properties being some sort of external IDs. Most URIs matched are simply the taxon Q-ID (Q16521). Of the unique URIs, ~30% are various Wikipedia links.
  • The total query time of taxon subgraph is ~3% of total query time. The average time per query is 0.064 seconds (64 ms). Almost all queries in this subgraph are small and simple.
  • The top user agent performs 85% of taxon subgraph queries, and the next 4 user agents combined perform 9% of the queries. The rest of the user-agents perform <1% of the queries in taxon subgraph. In terms of time, the top user agent, which made 85% of the queries, accounts for 33% of time consumed. But two more user agents perform some comparatively heavy queries. With less than 1% of queries each, they account for 26% and 11% of query time respectively.
  • The variety of query types in taxon subgraph is quite less (1.1K) compared to 11K query types in human subgraph. Only the top 3 types of queries account for ~90% of the queries of taxon subgraph. The rest form a long tail of small query counts.
  • Most user agents make only 1 type of query. Only 3 user agents make queries of >100 types, and 5 user agents make queries of 50-100 types. The top user agents (in terms of query time and count) make mostly <5 types of queries. Except for the two user agents that accounted for comparatively more query time, they perform 35-25 types of queries. Still quite a small number.
  • In sum, most queries in taxon subgraph are small and simple, take between 10 to 100ms time to run, and are mostly by 1/2 user agents.

What are subgraph related queries

We define some parameters to identify whether a query touches on a subgraph based on the items and properties a query uses. Some queries may even touch multiple subgraphs. See more on what a subgraph means here. Note: Subgraphs have overlaps.

The parameters that define which subgraph a query belongs to are:

1. If the query uses the subgraph's Q-ID. Example: Q5 containing queries are part of Q5 subgraph.
2. If the query uses items that are instance of a particular subgraph.
3. If the query uses items that occur 99% of the times in a particular subgraph.
4. If the query uses properties that occur 99% of the times in a particular subgraph.
5. If the query uses literals that occur 99% of the times in a particular subgraph. The literals can occur with or without language tags. Both versions are compared to check for match. Note that whole literals are matched in queries and Wikidata. Queries that ask for partial matches, using regex for example, are not included. The assumption is that such queries are more likely to contain other items from the subgraph and are caught anyways.

The following analysis uses Wikidata dump of 20211101 and WDQS public SPARQL queries of 10/2021 unless otherwise stated. All query related numbers below are monthly values.

Query count and time

• All queries here refer to queries with status code 200 and 500, i.e correct queries, successful or time-out.
• WDQS receives ~220M queries a month.
• Total query time for all queries for a month is ~16,000 hours.

The table below lists the top 50 most queried subgraphs with subgraph size and query time information of 11/2021. A breakdown of what caused the match is also present, which corresponds to the parameters mentioned in #What are subgraph related queries. It also ranks the subgraphs by size, query count, and query time consumed. A more complete list containing 341 subgraphs, that form ~90% of Wikidata triples, is available here: subgraph data for November 2021, and subgraph data for October 2021. The difference between values from October and November is shown in the next table for comparison purposes. In some places, the query count percentages differ slightly.

Top 50 most queries subgraphs in Wikidata with subgraph size information

Subgraph rank by size	Subgraph rank by query count	Subgraph rank by query time	Subgraph	Subgraph label	%of triples	%of entities	Days to recover (4.77M rate)	Query count	%count of all queries	Query time (hr)	%time of all queries	Avg time/query (s)	%query using only this subgraph	%query using this and other subgraphs	%count of query from Qid	%count of query from instance items	%count of query from items	%count of query from properties	%count of query from literals
3	1	1	Q5	human	7.254	10.045	204	60,868,572	31.941	5248	34.195	0.31	80.718	19.282	2.541	18.199	12.198	19.457	1.435
5	2	11	Q16521	taxon	2.885	3.5	81	27,172,995	14.259	480	3.131	0.064	95.179	4.821	12.19	0.746	12.862	0.87	0.433
34	3	7	Q4830453	business	0.107	0.208	3	9,228,037	4.842	554	3.607	0.216	68.097	31.903	1.646	2.95	2.24	0.001	0.177
6	4	5	Q101352	family name	1.646	0.511	46	5,990,617	3.144	659	4.292	0.396	20.867	79.133	0.041	3.057	2.791	0.018	0.038
15	5	2	Q11424	film	0.359	0.284	10	5,067,305	2.659	1541	10.042	1.095	74.15	25.85	0.451	1.469	1.348	0.003	0.543
1	6	13	Q13442814	scholarly article	48.935	39.815	1378	4,944,995	2.595	263	1.713	0.191	76.272	23.728	0.017	1.942	1.938	0.405	0.396
7	7	3	Q4167410	Wikimedia disambiguation page	1.354	1.464	38	3,292,873	1.728	765	4.982	0.836	22.922	77.078	0.164	0.192	0.472	0.0	1.163
2	8	25	Q6999	astronomical object	8.684	8.943	245	2,444,109	1.283	79	0.516	0.117	92.622	7.378	0.003	1.218	1.222	0.023	0.004
92	9	14	Q6881511	enterprise	0.036	0.052	1	1,937,486	1.017	234	1.528	0.436	2.584	97.416	0.083	0.812	0.538	0.0	0.071
26	10	29	Q484170	commune of France	0.179	0.048	5	1,934,902	1.015	70	0.455	0.13	78.322	21.678	0.024	0.869	0.085	0.115	0.01
19	11	22	Q13406463	Wikimedia list article	0.249	0.355	7	1,766,742	0.927	117	0.765	0.239	17.64	82.36	0.034	0.372	0.628	0.0	0.137
63	12	12	Q5398426	television series	0.055	0.063	2	1,379,486	0.724	411	2.68	1.073	66.963	33.037	0.048	0.376	0.369	0.0	0.167
37	13	47	Q7725634	literary work	0.087	0.203	2	1,377,546	0.723	42	0.273	0.11	6.75	93.25	0.39	0.181	0.243	0.0	0.009
16	14	4	Q486972	human settlement	0.298	0.612	8	1,328,064	0.697	699	4.557	1.896	42.354	57.646	0.328	0.39	0.236	0.0	0.005
163	15	15	Q891723	public company	0.015	0.013	0	1,175,813	0.617	219	1.426	0.67	10.916	89.084	0.042	0.415	0.185	0.001	0.092
90	16	6	Q43229	organization	0.037	0.082	1	1,067,340	0.56	600	3.908	2.023	40.99	59.01	0.259	0.227	0.146	0.0	0.021
13	17	24	Q3305213	painting	0.426	0.579	12	926,701	0.486	86	0.558	0.333	24.622	75.378	0.017	0.426	0.284	0.002	0.008
87	18	36	Q47461344	written work	0.037	0.078	1	881,216	0.462	53	0.345	0.216	4.058	95.942	0.289	0.079	0.114	0.0	0.003
25	19	32	Q532	village	0.199	0.294	6	872,310	0.458	61	0.399	0.253	39.133	60.867	0.003	0.417	0.198	0.0	0.015
4	20	28	Q4167836	Wikimedia category	5.806	5.175	164	808,536	0.424	74	0.484	0.331	81.351	18.649	0.037	0.363	0.292	0.0	0.024
61	21	51	Q7889	video game	0.055	0.048	2	753,351	0.395	37	0.244	0.179	62.267	37.733	0.006	0.181	0.314	0.002	0.01
20	22	41	Q8502	mountain	0.248	0.559	7	749,283	0.393	47	0.306	0.225	67.13	32.87	0.002	0.369	0.351	0.0	0.001
28	23	33	Q482994	album	0.16	0.288	5	704,746	0.37	59	0.388	0.304	27.474	72.526	0.012	0.15	0.189	0.0	0.098
89	24	17	Q4164871	position	0.037	0.128	1	645,434	0.339	175	1.141	0.977	12.504	87.496	0.003	0.305	0.025	0.0	0.011
8	25	16	Q7187	gene	0.911	1.273	26	604,364	0.317	208	1.354	1.238	31.577	68.423	0.084	0.1	0.022	0.015	0.127
11	26	26	Q11173	chemical compound	0.684	1.302	19	588,469	0.309	76	0.496	0.466	68.901	31.099	0.135	0.11	0.092	0.002	0.014
55	27	54	Q215380	musical group	0.062	0.087	2	585,266	0.307	37	0.241	0.227	39.016	60.984	0.01	0.205	0.16	0.0	0.011
31	28	39	Q16970	church building	0.128	0.227	4	577,677	0.303	48	0.315	0.301	43.769	56.231	0.003	0.288	0.214	0.0	0.002
71	29	55	Q732577	publication	0.047	0.076	1	569,536	0.299	37	0.238	0.231	4.203	95.797	0.283	0.015	0.296	0.0	0.0
22	30	43	Q79007	street	0.23	0.626	6	535,623	0.281	44	0.289	0.298	47.589	52.411	0.028	0.246	0.218	0.001	0.001
23	31	34	Q4022	river	0.216	0.425	6	520,347	0.273	56	0.365	0.388	53.592	46.408	0.002	0.254	0.192	0.0	0.002
242	32	8	Q14204246	Wikimedia project page	0.008	0.033	0	498,708	0.262	548	3.572	3.957	8.77	91.23	0.026	0.19	0.038	0.0	0.064
36	33	63	Q3947	house	0.096	0.216	3	465,249	0.244	33	0.212	0.252	58.051	41.949	0.0	0.238	0.223	0.0	0.002
32	34	31	Q41176	building	0.124	0.29	3	463,636	0.243	65	0.423	0.504	37.511	62.489	0.042	0.189	0.168	0.001	0.002
307	35	62	Q783794	company	0.005	0.012	0	459,638	0.241	33	0.213	0.256	44.132	55.868	0.081	0.146	0.1	0.0	0.006
29	36	48	Q23397	lake	0.136	0.279	4	456,054	0.239	42	0.273	0.331	59.859	40.141	0.002	0.227	0.211	0.0	0.001
119	37	42	Q3957	town	0.023	0.015	1	450,870	0.237	46	0.297	0.364	44.245	55.755	0.057	0.162	0.034	0.0	0.003
64	38	40	Q811979	architectural structure	0.054	0.12	2	445,779	0.234	48	0.313	0.388	12.038	87.962	0.097	0.126	0.117	0.0	0.001
80	39	59	Q34442	road	0.041	0.073	1	440,960	0.231	34	0.22	0.276	14.176	85.824	0.008	0.129	0.171	0.0	0.001
275	40	180	Q21198342	manga series	0.007	0.015	0	437,382	0.23	11	0.074	0.093	28.665	71.335	0.01	0.052	0.2	0.0	0.003
72	41	23	Q86850539	Whitaker's Latin frequency type C	0.047	0.011	1	436,103	0.229	95	0.622	0.788	10.35	89.65	0.0	0.0	0.0	0.0	0.228
138	42	139	Q18340514	events in a specific year or time period	0.019	0.048	1	431,649	0.227	16	0.104	0.133	10.729	89.271	0.0	0.21	0.068	0.0	0.004
261	43	53	Q2085381	publisher	0.007	0.015	0	420,459	0.221	37	0.243	0.319	52.906	47.094	0.001	0.21	0.068	0.0	0.004
44	44	38	Q55488	railway station	0.074	0.104	2	410,774	0.216	49	0.319	0.43	25.81	74.19	0.001	0.172	0.163	0.0	0.002
108	45	27	Q33506	museum	0.027	0.044	1	409,716	0.215	75	0.486	0.655	28.194	71.806	0.017	0.184	0.134	0.0	0.001
181	46	19	Q34770	language	0.013	0.011	0	402,013	0.211	145	0.947	1.302	33.166	66.834	0.009	0.169	0.02	0.0	0.017
112	47	86	Q15632617	fictional human	0.025	0.056	1	395,934	0.208	25	0.166	0.232	17.231	82.769	0.007	0.138	0.09	0.0	0.004
42	48	119	Q22808320	Wikimedia human name disambiguation page	0.077	0.075	2	381,873	0.2	19	0.125	0.181	67.093	32.907	0.0	0.164	0.142	0.0	0.001
143	49	75	Q11032	newspaper	0.017	0.043	0	380,153	0.199	28	0.181	0.263	55.697	44.303	0.002	0.169	0.143	0.0	0.019
38	50	117	Q3331189	version, edition, or translation	0.087	0.191	2	374,597	0.197	19	0.126	0.186	10.191	89.809	0.117	0.037	0.134	0.0	0.038

Comparison of subgraph queries across time

Comparison of subgraph queries across time (Oct, Nov 2021)

Subgraph rank by size	Subgraph	Subgraph label	%of entities	%of triples	Oct query count	Oct %count of queries	Oct query time (hr)	Oct %time of queries	Nov query count	Nov %count of queries	Nov query time (hr)	Nov %time of queries
3	Q5	human	9.986	7.324	68,659,369	31.058	6,314	39.3	60,868,572	31.941	5,248	34.195
5	Q16521	taxon	3.427	2.871	56,437,140	25.529	495	3.1	27,172,995	14.259	480	3.131
34	Q4830453	business	0.207	0.108	4,041,395	1.828	343	2.1	9,228,037	4.842	554	3.607
6	Q101352	family name	0.509	1.546	5,564,173	2.517	640	4.0	5,990,617	3.144	659	4.292
15	Q11424	film	0.281	0.364	4,757,084	2.152	1,613	10.0	5,067,305	2.659	1,541	10.042
1	Q13442814	scholarly article	39.794	49.668	1,649,268	0.746	142	0.9	4,944,995	2.595	263	1.713
7	Q4167410	Wikimedia disambiguation page	1.459	1.374	3,737,550	1.691	223	1.4	3,292,873	1.728	765	4.982
2	Q6999	astronomical object	8.942	8.75	448,032	0.203	51	0.3	2,444,109	1.283	79	0.516
92	Q6881511	enterprise	0.052	0.036	943,613	0.427	164	1.0	1,937,486	1.017	234	1.528
26	Q484170	commune of France	0.043	0.18	866,766	0.392	46	0.3	1,934,902	1.015	70	0.455
20	Q13406463	Wikimedia list article	0.352	0.252	1,283,160	0.58	73	0.5	1,766,742	0.927	117	0.765
63	Q5398426	television series	0.062	0.055	1,206,285	0.546	366	2.3	1,379,486	0.724	411	2.68
42	Q7725634	literary work	0.176	0.077	468,204	0.212	22	0.1	1,377,546	0.723	42	0.273
16	Q486972	human settlement	0.602	0.302	721,789	0.327	73	0.5	1,328,064	0.697	699	4.557
165	Q891723	public company	0.013	0.015	837,595	0.379	157	1.0	1,175,813	0.617	219	1.426
91	Q43229	organization	0.08	0.037	806,840	0.365	123	0.8	1,067,340	0.56	600	3.908
12	Q3305213	painting	0.578	0.432	834,752	0.378	79	0.5	926,701	0.486	86	0.558
86	Q47461344	written work	0.078	0.038	774,947	0.351	67	0.4	881,216	0.462	53	0.345
25	Q532	village	0.292	0.201	584,789	0.265	21	0.1	872,310	0.458	61	0.399
4	Q4167836	Wikimedia category	5.165	5.85	1,383,343	0.626	96	0.6	808,536	0.424	74	0.484
62	Q7889	video game	0.047	0.056	741,401	0.335	30	0.2	753,351	0.395	37	0.244
19	Q8502	mountain	0.559	0.253	227,393	0.103	16	0.1	749,283	0.393	47	0.306
28	Q482994	album	0.287	0.161	776,845	0.351	37	0.2	704,746	0.37	59	0.388
89	Q4164871	position	0.128	0.037	788,077	0.356	332	2.1	645,434	0.339	175	1.141
8	Q7187	gene	1.273	0.927	628,916	0.284	94	0.6	604,364	0.317	208	1.354
10	Q11173	chemical compound	1.302	0.693	1,307,852	0.592	133	0.8	588,469	0.309	76	0.496
54	Q215380	musical group	0.087	0.063	461,181	0.209	17	0.1	585,266	0.307	37	0.241
31	Q16970	church building	0.226	0.129	396,936	0.18	25	0.2	577,677	0.303	48	0.315
70	Q732577	publication	0.076	0.048	512,416	0.232	53	0.3	569,536	0.299	37	0.238
22	Q79007	street	0.62	0.231	225,188	0.102	20	0.1	535,623	0.281	44	0.289
23	Q4022	river	0.425	0.219	280,190	0.127	20	0.1	520,347	0.273	56	0.365
243	Q14204246	Wikimedia project page	0.033	0.008	1,114,113	0.504	62	0.4	498,708	0.262	548	3.572
36	Q3947	house	0.216	0.098	118,886	0.054	9	0.1	465,249	0.244	33	0.212
32	Q41176	building	0.287	0.125	271,666	0.123	36	0.2	463,636	0.243	65	0.423
310	Q783794	company	0.012	0.005	124,932	0.057	19	0.1	459,638	0.241	33	0.213
29	Q23397	lake	0.278	0.138	130,027	0.059	14	0.1	456,054	0.239	42	0.273
121	Q3957	town	0.015	0.023	294,685	0.133	24	0.1	450,870	0.237	46	0.297
64	Q811979	architectural structure	0.119	0.055	282,739	0.128	28	0.2	445,779	0.234	48	0.313
80	Q34442	road	0.073	0.041	215,771	0.098	14	0.1	440,960	0.231	34	0.22
280	Q21198342	manga series	0.014	0.007	208,503	0.094	5	0.0	437,382	0.23	11	0.074
71	Q86850539	Whitaker's Latin frequency type C	0.011	0.048	355,247	0.161	56	0.3	436,103	0.229	95	0.622
138	Q18340514	events in a specific year or time period	0.048	0.019	463,683	0.21	17	0.1	431,649	0.227	16	0.104
264	Q2085381	publisher	0.014	0.007	179,442	0.081	23	0.1	420,459	0.221	37	0.243
45	Q55488	railway station	0.104	0.075	258,862	0.117	20	0.1	410,774	0.216	49	0.319
108	Q33506	museum	0.044	0.028	252,308	0.114	54	0.3	409,716	0.215	75	0.486
177	Q34770	language	0.011	0.013	1,713,196	0.775	73	0.5	402,013	0.211	145	0.947
113	Q15632617	fictional human	0.056	0.026	306,319	0.139	18	0.1	395,934	0.208	25	0.166
41	Q22808320	Wikimedia human name disambiguation page	0.075	0.078	433,986	0.196	17	0.1	381,873	0.2	19	0.125
144	Q11032	newspaper	0.043	0.017	230,085	0.104	11	0.1	380,153	0.199	28	0.181
37	Q3331189	version, edition, or translation	0.19	0.087	410,352	0.186	34	0.2	374,597	0.197	19	0.126

More on query time

The query time can be broken down to classes for better visualization. Below is a figure with the query class distribution (number of queries per query time class per subgraph) for the top 50 subgraphs w.r.t query time consumed. Some of the takeaways are:

• Most subgraphs have most queries in the range of 10ms to 100ms
• Second most common class is 100ms to 1s
• collection and photograph have most queries (~150k) timed at 1s to 10s. Around 10 more subgraphs have a little (~10-20k) query in this time range.

File:Top 50 query time class.png

Distribution comparison

Following is the query time distribution of all queries, regardless of subgraph.

We then compare this distribution to the distributions listed above for the top subgraphs. To compare, we plot the differences of percentages in each subgraph with the percentage in all queries. That is,

For each subgraph:
 For each query time class:
  Percent of query count for this time class in this subgraph - Percent of query count in all queries

This shows us the difference in overall distribution with the distribution in the individual subgraphs.

File:Top 50 time class dist diffs.png

User agent

Analysis on user-agent is an approximation because these don't completely represent distinct users. For example lots people use the same bot or script without changing the user-agent, or the same person or bot uses multiple user-agent strings. Yet based on the available data we can get an estimate nevertheless.

User agent count

• Total number of unique user agents across all subgraphs: 981,180
• First, a list of subgraphs with most and least distinct user-agents is listed. It seems the least number of user-agents a subgraph has is at least 10. So the large subgraphs are used by at least a bunch of users.
• The largest numbers of user-agents are present in a variety of type of subgraphs. gene, protein, biological process, molecular function appear to be similar among them. It is possible the same queries represent several of these subgraphs. More on subgraph connectivity in #Subgraph Connectivity.

Subgraphs with most user-agents Subgraph Subgraph label %Query #User agents %User agent Q11424 film 2.152 251,420 0.256 Q8054 protein 0.158 234,659 0.239 Q7187 gene 0.284 187,029 0.191 Q2996394 biological process 0.072 124,415 0.127 Q14860489 molecular function 0.044 89,445 0.091 Q5 human 31.058 55,377 0.056 Q898273 protein domain 0.019 38,484 0.039 Q16521 taxon 25.529 25,193 0.026 Q86850539 Whitaker's Latin frequency type C 0.161 20,158 0.021 Q4167410 Wikimedia disambiguation page 1.691 13,818 0.014 Q14204246 Wikimedia project page 0.504 13,443 0.014 Q476028 association football club 0.145 12,086 0.012 Q235557 file format 0.045 7,701 0.008 Q1520033 count noun 0.05 7,662 0.008 Q417841 protein family 0.007 4,906 0.005 Q484170 commune of France 0.392 4,764 0.005 Q4830453 business 1.828 4,383 0.004 Q4164871 position 0.356 4,319 0.004 Q7278 political party 0.109 4,073 0.004 Q3918 university 0.104 3,565 0.004

Subgraphs with least user-agents Subgraph Subgraph label %Query #User agents %User agent Q106006703 local regulations of the People's Republic of China 0.0 11 0.0 Q67015940 Government Boys' Primary School 0.0 13 0.0 Q7604693 Statutory Rules of Northern Ireland 0.0 13 0.0 Q106474968 ethnic group by settlement in Macedonia 0.003 15 0.0 Q6453643 decree law 0.0 15 0.0 Q97695005 committee group motion 0.0 15 0.0 Q100532807 Irish Statutory Instrument 0.0 16 0.0 Q10429085 report 0.0 19 0.0 Q99045339 written question 0.0 20 0.0 Q1505023 Interpellation 0.0 20 0.0 Q96739634 individual motion 0.0 21 0.0 Q67035425 ASTM standard 0.0 21 0.0 Q61278455 health sub-centre 0.001 23 0.0 Q26267864 Wikimedia KML file 0.005 23 0.0 Q3508250 Syndicat intercommunal 0.02 24 0.0 Q107102664 cell line from embryonic stem cells 0.0 24 0.0 Q7604686 UK Statutory Instrument 0.0 27 0.0 Q6451276 Congressional Research Service report 0.001 28 0.0 Q61443650 sub post office 0.0 33 0.0 Q26894053 basketball team season 0.009 34 0.0

• There are 50 subgraphs with more than 1000 user agents, and 300 subgraphs with less than 1000 user agents. Most subgraphs are therefore not queried by too many distinct users. The distribution of user-agent counts less than 1000 is shown in the figure below. This clearly shows the small number of user counts in most subgraphs.

User agent distribution in subgraphs

• Next, the user agent vs query count distribution was analyzed for some of the top subgraphs. While user agent count gives us an idea of how many users may be using a subgraph, it is not clear whether all of them query the subgraph equally, or very few user agents perform most of the queries.
• ~30 out of 341 subgraphs have a user agent that queries >=50% of all queries of that particular subgraphs.
• 6 subgraphs have a user agent querying around 80-90% of the time.
• So the trend of a dominating single source of queries is not wide spread among subgraphs, but is present in few subgraphs nonetheless.

The figure below shows the top 2 user-agent query in percents for 341 subgraphs. This shows whether there is a dominating pattern in a subgraph with the top user agents per subgraph. This figure shows the top 2 user-agent query percents for 341 subgraphs. This shows whether there is a dominating pattern in a subgraph with the top user agents per subgraph.

The figure below shows 100 subgraphs with their user agent query usage distribution in percents. Usage greater than 50% is marked in red. A birds-eye view of the plots shows how some subgraphs have a dominating user agent and most other subgraphs have at least 1 or 2 user agents that query the most. The rest of the user agents form the long tail of the distribution This figure shows 100 subgraphs with their user agent query usage distribution in percents. Usage greater than 50% is marked in red. A birds-eye view of the plots shows how some subgraphs have a dominating user agent while most subgraphs have at least 1 or 2 user agents that query the most. The rest of the user agents form the long tail 10% of the distribution.

Top user agents in subgraphs

• The top user agents in various subgraphs is listed below. More analysis on Q5 (human) and Q16521 (taxon) is done at the end of the page as they are the most queried subgraphs.

Top user agents in various subgraphs

Subgraph	Subgraph label	User agent	Query count (in subgraph)	Query percent (within subgraph)	Query percent overall
Q16521	taxon	mix-n-match	50,622,670	89.697	22.899
Q5	human	UA # 2	9,017,930	13.134	4.079
Q5	human	mix-n-match	8,548,335	12.45	3.867
Q5	human	UA # 3	5,059,258	7.369	2.289
Q5	human	UA # 4	4,020,496	5.856	1.819
Q5	human	UA # 5	3,828,747	5.576	1.732
Q101352	family name	UA # 5	3,828,747	68.811	1.732
Q5	human	UA # 6	2,685,807	3.912	1.215
Q5	human	UA # 7	2,434,486	3.546	1.101
Q4830453	business	UA # 8	2,403,677	59.476	1.087
Q5	human	UA # 9	2,020,598	2.943	0.914
Q16521	taxon	Hub	1,984,437	3.516	0.898
Q5	human	UA # 11	1,877,700	2.735	0.849
Q5	human	UA # 12	1,781,161	2.594	0.806
Q16521	taxon	UA # 13	1,294,113	2.293	0.585

User agent vs Subgraph

So far we have explored the user-agent count and distribution per subgraph. It is also important to note the user agent's query across subgraphs. In other words,

• Do users have a very specific use case and so the queries spans only a few subgraphs? or is it spread across a lot of subgraphs?
• Are there some user agents that query the most in multiple subgraphs? This could be due to the nature of the use case or simply because some subgraphs overlap a lot.

We start by looking at how many user agents access how many subgraphs. From the table below, we see that most user agents (89% of them) query one subgraph only. Some user agents query a lot of subgraphs as well. A clearer picture is seem from the plot below.

Relationship between subgraphs and user agents #of Subgraphs (X) #of User agents querying X subgraphs %of User agents querying X subgraphs 1 875,724 89.252 2 91,962 9.373 5 3,562 0.363 3 2,388 0.243 6 1,539 0.157 7 799 0.081 9 628 0.064 8 463 0.047 4 460 0.047 12 332 0.034 16 308 0.031 15 282 0.029 10 281 0.029 17 242 0.025 18 235 0.024 14 202 0.021 11 184 0.019 19 177 0.018 13 167 0.017 20 119 0.012 21 75 0.008 22 47 0.005 25 46 0.005 23 39 0.004 24 39 0.004 27 32 0.003 26 28 0.003 28 26 0.003 29 25 0.003 30 20 0.002 31 17 0.002 35 16 0.002 37 16 0.002 47 15 0.002 34 15 0.002 61 13 0.001 32 12 0.001 50 12 0.001 36 11 0.001 44 11 0.001 49 10 0.001 65 9 0.001 56 9 0.001 72 9 0.001 51 9 0.001 121 9 0.001 95 9 0.001 124 9 0.001 42 9 0.001 39 9 0.001

File:Ua vs subgraph.png

Next we isolate user agents from each subgraph who query drastically more (>=10% difference) than other user agents in the same subgraph, and perform at least 100k queries (0.05% of all queries) a month. A list of ~30 such user agents was found. A plot with subgraph distributions of all these user agents was observed to find some large buckets where they tend to query. The plot is shows below, followed by some explicit observations.

Percentages below are percent of all monthly queries. mix n match (UA #17): a lot of taxon queries (Q16521), 23% a lot of human queries (Q5), 4% UA #6: 1% in Business (Q4830453) UA #14: 1% in human (Q5) 0.5% in film (Q11424) UA #23: 1.73% in family name (Q101352) 1.73% in human (Q5) both have exact counts, meaning they could be the same queries that touch both human and family name subgraphs

For reference: 100% percent is 221,067,674 queries 10% percent is 22,106,767 queries 1% percent is 2,210,676 queries 0.1% percent is 221,067 queries 0.05% percent is 110,533 queries 0.01% percent is 22,106 queries

Subgraph connectivity through queries

Subgraph connectivity was explored to some extent using only Wikidata in Wikidata_Subgraph_Analysis. This was based on what items or properties were common between subgraphs and how many direct connections were present between them. A visualization was created to show the strength of this connectivity between subgraphs here: wikidata_graph. This section aims to analyze the connectivity of subgraphs through the queries, i.e, how often are some subgraphs queried together.

• Subgraph Queries: The total number of queries that touch at least one of the top 341 subgraphs is 72% of all queries.
• First we look at how many subgraphs do most queries access. The tables below show the least and most query groups by number of subgraphs accessed.
• 70% of all queries (97% of subgraph queries) touch on 1 or 2 subgraph. 64% of all queries (90% of subgraph queries) touch on only 1 subgraph.

Queries with most subgraphs accessed #of Subgraphs #of Queries 341 25 333 1 315 2 313 3 258 1 181 3 152 1 142 1 133 2 130 2 129 1 128 2 127 4 126 4 125 9

Queries with least subgraphs accessed #of Subgraphs #of Queries %of Queries 1 142,507,736 64.463 2 12,464,811 5.638 3 1,767,253 0.799 4 586,173 0.265 5 364,445 0.165 6 221,485 0.1 7 188,012 0.085 8 112,922 0.051 9 102,524 0.046 10 68,871 0.031 11 50,341 0.023 12 38,102 0.017 13 34,075 0.015 14 24,003 0.011 15 17,935 0.008

• It is hard to view which subgraphs occur together from the data above. So the subgraphs that occurred together were broken into pairs and pars of subgraphs that occur together the most were listed.
• There are 57,970 subgraphs pairs that occur together in queries. Total possible subgraph pair count is (340*341)/2 = 57,970. This shows that every subgraph is connected to every other subgraph through queries! Of course the number of queries vary widely.
• A list of some of the most queried subgraph pairs is shown below.

Top pairs of subgraphs that are queried together

Subgraph 1		Subgraph 2		Query
Subgraph	Subgraph label	Subgraph	Subgraph label	#of Query	%of Query
Q101352	family name	Q5	human	4,649,345	2.44
Q4830453	business	Q6881511	enterprise	1,858,183	0.975
Q11424	film	Q5	human	1,096,150	0.575
Q5	human	Q7725634	literary work	1,067,191	0.56
Q4830453	business	Q891723	public company	973,565	0.511
Q13406463	Wikimedia list article	Q5	human	970,047	0.509
Q16521	taxon	Q5	human	890,304	0.467
Q4167410	Wikimedia disambiguation page	Q5	human	840,151	0.441
Q4830453	business	Q5	human	680,786	0.357
Q3305213	painting	Q4167410	Wikimedia disambiguation page	606,434	0.318
Q6881511	enterprise	Q891723	public company	572,986	0.301
Q13442814	scholarly article	Q5	human	527,538	0.277
Q47461344	written work	Q732577	publication	514,321	0.27
Q4164871	position	Q5	human	480,484	0.252
Q13442814	scholarly article	Q4167410	Wikimedia disambiguation page	446,490	0.234
Q482994	album	Q5	human	409,139	0.215
Q13406463	Wikimedia list article	Q16521	taxon	401,466	0.211
Q13406463	Wikimedia list article	Q4167410	Wikimedia disambiguation page	349,421	0.183
Q14204246	Wikimedia project page	Q4167410	Wikimedia disambiguation page	341,845	0.179
Q43229	organization	Q5	human	337,868	0.177
Q5398426	television series	Q5	human	323,501	0.17
Q215380	musical group	Q5	human	320,532	0.168
Q47461344	written work	Q5	human	313149	0.164
Q5	human	Q6881511	enterprise	285,110	0.15
Q3331189	version, edition, or translation	Q5	human	283,741	0.149
Q5	human	Q86850539	Whitaker's Latin frequency type C	280,866	0.147
Q11424	film	Q13406463	Wikimedia list article	272,316	0.143
Q13406463	Wikimedia list article	Q18340514	events in a specific year or time period	270,710	0.142
Q16521	taxon	Q4167410	Wikimedia disambiguation page	266,507	0.14
Q4167410	Wikimedia disambiguation page	Q86850539	Whitaker's Latin frequency type C	249,340	0.131

• The distribution of the number of times each subgraph pair in Wikidata occurs in queries is shown below. Note that (A,B) pair is the same as (B,A) pair, so there is no duplication in the plots. Since the plot is extremely skewed, three plots with various limits on the number of occurrences are shown. We can see how only a small number of pairs occur a lot together, they can be viewed from the table above. Whereas a huge number of pairs occur a very small number of times.

• Below is a heat-map of the number of queries, where both x and y axis represent subgraph indices (names of subgraphs not shown due to space). The subgraphs are sorted by most queried subgraphs.
• The diagonals show queries that use only 1 subgraph and are represented as Q5-Q5, or Q42-Q42 for example. Other are represented as Q5-Q42 or Q42-Q5
• It is a Symmetrical plot.
• The tons of vertical and horizontal lines indicate there are lots of subgraphs that happen to pair with many other subgraphs.
• The exact numbers for this heat-map are present in subgraph_pair_heatmap_df as a csv file.

Number of subgraph accessed vs query time

To view whether there is a correlation between accessing more subgraphs and query time, various subsets of subgraphs were taken and their query time distributions were observed. Then, number (and percents) of queries that access various number of subgraphs were plotted for each query time group. A simple scatter plot with time and subgraph number was not possible due to the large number of queries, but the following plots give us a good idea of the correlation. The pearson correlation of query time and number of subgraphs accessed is 0.016. We see there is a slight correlation but it is not significant enough. All query time groups are dominated by queries that access 1 or 2 subgraphs. Queries accessing more subgraphs do appear comparatively more in More than 10s group.

The following analysis was done with data from November 2021. Thus there are slight differences in numbers from the above analysis, which were done with October 2021 data.

Paired subgraph query time analysis

We observed that some queries do indeed take more time when more subgraphs are involved. But this could also occur because of the particular subgraphs being accessed. For example, even simple queries in the scholarly article subgraph may take a long time time out simply due to the large size of graph it has to comb through. To make this analysis complete, we look at the query times of queries that access only subgraph X and those that access X and other subgraphs. The influence of the other subgraphs persists, but we can now pick out anything clearly similar or different. If both plots look similar (the query classes), then we can assume it to be the effect of subgraph X. If there are more long-running queries when the query accesses X with other subgraphs, than when it accesses only X, then we can assume the cause of this long-running queries is not solely X. It could be due to other subgraphs, or simply due to the nature of the query (too complex, lots of string manipulation, regex etc).

The plots below shows the comparison of query time class distribution for when the queries use only subgraph X versus when they use X and other subgraphs, where X represents the top 30 large subgraphs. It indeed looks like some subgraphs have queries with longer time when other subgraphs are involved, such as for scholarly articles, lake, hill, clinical trial, river etc.

Human subgraph (Q5) query analysis

The following analysis was done with query data of November, 2021.

The queries that were estimated to be related to the human subgraph accounted for 31.94% of all queries in Wikidata. 25.78% queries used only the human subgraph and the rest 6.16% queries used a mix of human and various other subgraphs. As described in #What are subgraph related queries, subgraphs are related to queries through Properties, Subject or Object URIs, Subgraph instance items, etc. Here is a breakdown for human subgraph taken from #Query count and time. A query can be said to be related to human subgraph due to multiple of the following reasons.

• Number of queries: 60,868,572 (31.94%)
• Percent of queries matching subgraph Qid, i.e, has Q5: 2.54%
• Percent of queries matching instance items: 18%
• Percent of queries matching subject/object URIs: 12%
• Percent of queries matching properties: 19.45%
• Percent of queries matching literal strings: 1.43%

Some of these breakdown have large percentages. It is worth looking at what items/properties/URIs are queried the most. Also looking at the distribution of such items' usage in queries shows how narrow or wide the search space is.

Here is a detailed breakdown of what kind of match caused a query to be part of the human subgraph:

Human subgraph query breakdown item predicate URI human Q-id literal # query % all query % human query 0 1 0 0 0 17,785,347 9.333 29.219 1 1 1 0 0 12,215,379 6.41 20.068 1 0 0 0 0 10,705,360 5.618 17.588 1 0 1 0 0 7,253,287 3.806 11.916 1 1 0 0 0 3,137,130 1.646 5.154 0 0 1 0 0 2,512,142 1.318 4.127 0 0 0 0 1 1,775,347 0.932 2.917 0 1 0 1 0 1,694,236 0.889 2.783 0 0 0 1 0 930,137 0.488 1.528 1 1 0 1 0 598,261 0.314 0.983 1 1 1 1 0 508,706 0.267 0.836 0 1 0 0 1 407,610 0.214 0.67 0 0 0 1 1 350,982 0.184 0.577 0 1 1 1 0 311,340 0.163 0.511 0 1 1 0 0 226,959 0.119 0.373 1 0 0 1 0 178,650 0.094 0.294 0 1 1 1 1 135,684 0.071 0.223 1 0 1 1 0 76,736 0.04 0.126 0 1 0 1 1 56,971 0.03 0.094 1 0 1 0 1 3,451 0.002 0.006 1 0 0 0 1 2,844 0.001 0.005 0 0 1 0 1 702 0.0 0.001 1 1 1 1 1 437 0.0 0.001 1 1 0 1 1 393 0.0 0.001 0 0 1 1 0 304 0.0 0.0 1 1 0 0 1 93 0.0 0.0 1 0 0 1 1 59 0.0 0.0 1 1 1 0 1 17 0.0 0.0 1 0 1 1 1 5 0.0 0.0 0 1 1 0 1 3 0.0 0.0 Total 60,868,572 31.94 100

Instance items matched

• Total items used: 7,969,182
• Total queries that use these items: 34,680,808 (18% of all queries)
• The distribution shows there are some high usage (~10k-20k queries) items, a small number of medium usage (~5k queries) items, and rest form a long tail of small usage (<1k queries) items in the human subgraph.

Top items that cause a query to be related to Human subgraph (Q5) Instance item Instance item label #of queries Q22686 Donald Trump 19,759 Q1747297 Robert Oliveri 19,247 Q509260 John Zimmerman 19,193 Q6499255 Laura Nader 19,135 Q209394 Michael Wood 19,101 Q937 Albert Einstein 19,098 Q7340648 Rob Whitehurst 19,026 Q52354375 Irene Aparicio 18,970 Q6232209 John F. Cassidy 18,964 Q22986632 Lori Lynn Ross 18,954 Q3976229 Stuart Lancaster 18,953 Q106466114 Gary Michael Ritchie 18,947 Q86599148 James Spicer 18,926 Q87653156 David A. Cook 18,919 Q16015822 Jerry Fleck 18917 Q7179427 Petur Hliddal 18,914 Q19878977 Jackie Carson 18,902 Q99859767 Kathy McCarty 18,898 Q90307934 Ann Harris 18,893 Q1070508 Cheryl Carasik 18,834 Q9682 Elizabeth II 18,816 Q6279 Joe Biden 18,277 Q64840837 Dylan Arnold 18,161 Q76 Barack Obama 18,035 Q107626126 Mauricio Lara 18,010

Properties matched

• Total properties used: 1,091 (Recall these are properties that occur 99% of the times in the human subgraph)
• Total queries that use these properties: 37,078,566 (19.45% of all queries)
• The distribution shows there are 3 properties with ~20-30M queries, 7 properties with ~1-5M queries, and rest of the more than 1000 properties match ~100K and less queries. In short, the distribution is a extremely skewed by only ~10 properties that are highly related to the human subgraph.

Top properties that cause a query to be related to Human subgraph (Q5) Property Property label #of queries P570 date of death 30,151,024 P569 date of birth 30,084,200 P27 country of citizenship 24,186,000 P106 occupation 5,259,920 P734 family name 4,871,326 P735 given name 4,616,631 P19 place of birth 2,379,702 P2949 WikiTree person ID 1,707,373 P20 place of death 1,222,037 P4985 TMDb person ID 916,399 P39 position held 750,380 P3602 candidacy in election 599,067 P69 educated at 561,380 P26 spouse 471,589 P108 employer 384,111 P2562 married name 279,197 P937 work location 258,707 P1066 student of 158,339 P184 doctoral advisor 152,318 P1960 Google Scholar author ID 151,507 P185 doctoral student 150,982 P54 member of sports team 150,573 P1153 Scopus author ID 150,545 P119 place of burial 144,027 P3829 Publons author ID 138,839

Subject/Object URI matched

• Total URIs used: 7,926,297 (Recall these are URIs that occur 99% of the times in the human subgraph)
• Total queries that use these URIs: 23,245,152 (12.2% of all queries)
• The top URIs/items show the obvious and most common ways the human subgraph is queried: query about specific people, about groups of people, and about their wikipedia pages. More about types of queries below.
• The distribution is a smooth logarithmic graph with only one item present in 165k queries, and the rest go down from 40k in a logarithmic pattern.

Top URIs that cause a query to be related to Human subgraph (Q5) URI URI label #of queries Q3391743 visual artist 165,540 Q1925963 graphic artist 38,897 Q28389 screenwriter 33,718 en.wikipedia.org/wiki/Lee_Child - 33,179 en.wikipedia.org/wiki/Emily_Wilson_(journalist) - 30,837 en.wikipedia.org/wiki/M.I.A._(rapper) - 29,388 Q10800557 film actor 29,318 en.wikipedia.org/wiki/Shannon_Lee - 29,216 en.wikipedia.org/wiki/Eugene_Gordon_Lee - 29,205 en.wikipedia.org/wiki/Lee_Childs - 29,203 en.wikipedia.org/wiki/Emily_Wilson_(classicist) - 26,864 en.wikipedia.org/wiki/Emily_Wilson_(actress) - 26,862 en.wikipedia.org/wiki/Adhir_Kalyan - 26,862 en.wikipedia.org/wiki/Emily_Wilson_(footballer) - 26,861 en.wikipedia.org/wiki/Emily_Wilson_Walker - 26,861 Q10798782 television actor 24,679 Q185351 jurist 22,130 Q1650915 researcher 21,206 Q2374149 botanist 20,385 Q250867 Catholic priest 20,314 Q10873124 chess player 19,832 Q12299841 cricketer 19,414 Q14373094 rugby league player 19,396 Q509260 John Zimmerman 19,193 Q6499255 Laura Nader 19,135

Query time

• The total query time of human subgraph is 34% of total query time and total query count is ~32% of all queries.
• Average time per query is 0.3 seconds (300 ms). Most queries in this subgraph are small and simple.
• The query time distribution is shown in the chart below, both in absolute counts and in percent of queries in human subgraph.

User agent

List of top user agents that query human subgraph is given below. This helps us view the distribution of usage - whether few user agents dominate the usage or it is a rather well distributed usage scenario across user agents. Top 10 user agents in terms of query count and query time is shown in the table below.

Top user agents in human subgraph

User agent	Query count	% query in human subgraph	% query overall	Query time(hr)	% query time in human subgraph	% query time overall
mix-n-match	6,960,988	11.436	3.653	79	1.51	0.516
searx1	6,615,319	10.868	3.471	778	14.832	5.072
UA#3	3,491,821	5.737	1.832	75	1.426	0.487
UA#4	3,073,725	5.05	1.613	175	3.327	1.138
UA#5	2,933,240	4.819	1.539	80	1.516	0.518
UA#6	2,488,807	4.089	1.306	19	0.364	0.125
UA#7	2,182,220	3.585	1.145	44	0.841	0.288
WikidataQueryServiceR	2,044,045	3.358	1.073	36	0.68	0.232
UA#9	1,970,264	3.237	1.034	27	0.524	0.179
searx2	1,909,144	3.137	1.002	200	3.808	1.302
UA#11	75,523	0.124	0.04	434	8.271	2.828
UA#12	55,357	0.091	0.029	319	6.083	2.08
searx3	1,428,789	2.347	0.75	151	2.871	0.982
OB-bot	287,534	0.472	0.151	144	2.736	0.935
UA#15	50,915	0.084	0.027	134	2.553	0.873
UA#16	31,298	0.051	0.016	112	2.132	0.729
searx4	771,932	1.268	0.405	92	1.761	0.602

The query time breakdown was plotted for the top 20 user agents (in terms of time). Most queries have query time of 10ms to 1s, as observed earlier. Some user agents have most queries in the range 10ms to 100ms and some others have most queries in the range 100ms to 1s.

Query types

Query types are grouped by the operations a query uses and also the order of operations used. This groups similar queries together despite different information sought and also separates groups of simple or complicated queries. The human subgraph has ~11,500 different types of queries. Notice that some query groups can be very similar in what they ask for, while most groups differ a lot. The top query groups are listed below. The top 10 types of queries account for 60% of the queries of human subgraph. The rest form a really long tail of small query counts.

Top query types in human subgraph

Query group (operation list)	#of queries	%of queries in human subgraph
bgp, service, path, bgp, sequence, leftjoin, bgp, join, path, bgp, sequence, leftjoin, bgp, join, path, bgp, sequence, leftjoin, bgp, join, path, bgp, sequence, leftjoin, bgp, join, path, bgp, sequence, leftjoin, bgp, join, path, bgp, sequence, leftjoin, bgp, join, path, bgp, sequence, leftjoin, bgp, join, path, bgp, sequence, leftjoin, bgp, join, bgp, (leftjoin, bgp)x10, path, leftjoin, leftjoin, bgp, path, leftjoin, leftjoin, bgp, leftjoin, bgp, path, leftjoin, (leftjoin, bgp)x15, leftjoin, path, bgp, sequence, leftjoin, bgp, join, bgp, (leftjoin, bgp)x35, leftjoin, path, bgp, sequence, (leftjoin, bgp)x8, service, join, group, (extend)x68,project	9,511,204	15.626
table, bgp, join, filter, project, distinct	6,889,023	11.318
bgp, service, path, bgp, sequence, leftjoin, bgp, join, path, bgp, sequence, leftjoin, bgp, join, path, bgp, sequence, leftjoin, bgp, join, path, bgp, sequence, leftjoin, bgp, join, path, bgp, sequence, leftjoin, bgp, join, path, bgp, sequence, leftjoin, bgp, join, path, bgp, sequence, leftjoin, bgp, join, path, bgp, sequence, leftjoin, bgp, join, bgp, (leftjoin, bgp)x9, leftjoin, bgp, path, leftjoin, leftjoin, bgp, path, leftjoin, leftjoin, bgp, leftjoin, bgp, path, leftjoin, (leftjoin, bgp)x15 leftjoin, path, bgp, sequence, leftjoin, bgp, join, bgp, (leftjoin, bgp)x36, leftjoin, path, bgp, sequence, (leftjoin, bgp)x8, service, join, group, (extend)x68, project	4,298,916	7.063
table, bgp, join, bgp, leftjoin, bgp, leftjoin, bgp, join, filter, project	3,444,363	5.659
bgp, bgp, leftjoin, filter, bgp, extend, filter, union, bgp, extend, filter, union, project]	3,073,725	5.05
bgp, project	2,454,919	4.033
table, bgp, leftjoin, bgp, join, filter, project	2,429,518	3.991
bgp, bgp, service, join, filter, project, distinct	1,172,912	1.927
table, bgp, join, bgp, service, join, order, project	1,047,351	1.721
table, extend, bgp, join, bgp, leftjoin, bgp, leftjoin, bgp, leftjoin, path, leftjoin, bgp, leftjoin, bgp, leftjoin, bgp, leftjoin, bgp, leftjoin, bgp, leftjoin, bgp, leftjoin, bgp, leftjoin, bgp, leftjoin, bgp, service, join, filter, project	1,033,220	1.697

Looking at the top 20 query groups (that make up 70% of all human subgraph queries), the following query types were found:

1. Query lists lots of predicates with times/date precision of some items
2. Mix n Match: Birth and Death of certain people, with filters
3. Like 1, slightly different
4. Name and family information of people in specific languages
5. All properties of some humans (these queries are generic but here used for humans)
6. Uses P227 (or some international ID) and retrieves the Wikipedia page for it
7. Same as 6 but written differently
8. Just wants labels in specific languages
9. Same as 6 but written differently
10. All contact info of people (facebook, instagram, youtube, twitter, etc)
11. Labels and Wikipedia article of people
12. Only label and description of people
13. Search for films by director filter
14. CEO or high officials of companies searched by name strings
15. Timeline of people of a particular occupation and particular gender
16. Occupation, name, birth, death of people
17. Education institution and its reference of people
18. Label and Wikipedia page of people in specific languages
19. List all people, or all people of certain occupation, or entity related to a given Wikipedia page (generic query)
20. Notable works, labels, and IDs of works like isbns

UA vs query types

Getting the number of query types per user agent informs us of the variety of queries a user agent makes to WDQS. This also breaks down the human subgraph queries into finer groups. The following plot shows the number of query types for each user agent in the human subgraph.

This shows us that most user agents make only 1 type of query. Only 8 user agents make queries of >500 types, and ~50 user agents make queries of >100 types. Looking into query counts in each of these UA - query type groups, we find that they have few queries (<10,000), and only ~10 groups have >10,000 queries, but all of these are small simple queries. The figure below shows the number of query per query type for the top 8 user agents. As we can see, their distribution looks alike although their query counts and query types differ.

Query type vs time class

While there are close to 11,500 query types, 20 of these types make 70% of all queries of human subgraph (22% of all queries), not all of them are equally time consuming. Some can be simple queries, while some can be long and complex. The following plot shows these 20 query types with query time classes. The values above the bar show both percent in human subgraph and overall query percentage. The subplots are titled with percent of the number of queries in that query type.

Services

The queries use ~50 unique services. The top 10 services are the most used; rest are used in less than 50 queries, mostly in less than 10 queries. 20 of these services are used in only 1 query.

Top 10 used services in queries in human subgraph

Service	Query count	% query in human subgraph
wikibase:label	29,925,798	49.165
wikibase:mwapi	14,014,458	23.024
gas:service	46,588	0.077
bd:slice	42,764	0.07
http://dbpedia.org/sparql	22,751	0.037
https://query.wikidata.org/sparql	22,751	0.037
wikibase:around	1,733	0.003
https://sophox.org/sparql	628	0.001
wikibase:box	195	0.0
mediawiki:categoryTree	45	0.0

Triples

Some query type analysis done in section query types gives us a good idea of what kind of queries human subgraph receives. Looking at the triples themselves also helps us peek into what most of the queries look like, what the most common subjects, objects, and properties are. The table below lists these along with the top Wikidata items and properties used overall. From the numbers it seems the top items are probably part of the same queries.

Top Subjects Subject Query count % Query count in human subgraph bd:serviceParam 30,022,819 49.324 item 23,632,602 38.826 hint:Prior 14,198,844 23.327 P27 13,845,887 22.747 P17 13,810,191 22.689 articleen 13,810,173 22.689 P281Node 13,810,142 22.688 P2048Node 13,810,142 22.688 P2046Node 13,810,142 22.688 P400 13,810,128 22.688 P275 13,810,128 22.688 P1346 13,810,128 22.688 P162 13,810,128 22.688 P35 13,810,128 22.688 P495 13,810,128 22.688 P112 13,810,128 22.688 P277 13,810,128 22.688 P123 13,810,128 22.688 P282 13,810,128 22.688 P58 13,810,128 22.688

Top Predicates Predicate Query count % Query count in human subgraph rdfs:label 34,247,253 56.264 wikibase:language 29,925,828 49.165 schema:about 24,610,856 40.433 schema:description 22,181,059 36.441 schema:isPartOf 20,672,719 33.963 schema:inLanguage 19,811,149 32.547 wdt:P27 18,580,218 30.525 wdt:P18 16,407,156 26.955 wdt:P50 15,283,299 25.109 wdt:P856 15,277,657 25.099 wdt:P2002 15,260,957 25.072 wdt:P345 15,175,422 24.931 rdf:type 15,141,091 24.875 wdt:P2013 15,126,300 24.851 wdt:P2003 15,044,080 24.716 wdt:P2397 14,980,826 24.612 wdt:P212 14,433,555 23.713 wikibase:timePrecision 14,224,353 23.369 wikibase:timeValue 14,222,729 23.366 wdt:P17 14,130,982 23.216

Top Objects Object Query count % Query count in human subgraph en 23,579,559 38.738 https://en.wikipedia.org/ 16,123,553 26.489 item 15,770,101 25.908 itemLabel 14,812,538 24.335 itemDescription 14,040,107 23.066 www.wikidata.org 14,008,069 23.014 true^^http://www.w3.org/2001/XMLSchema#boolean 13,994,657 22.992 mwapi:item 13,899,357 22.835 EntitySearch 13,897,505 22.832 P27 13,895,366 22.828 P18 13,859,909 22.77 wikibase:BestRank 13,825,048 22.713 P569 13,811,781 22.691 P570 13,811,013 22.69 1^^http://www.w3.org/2001/XMLSchema#integer 13,810,316 22.689 P17 13,810,214 22.689 P856 13,810,151 22.688 P50 13,810,144 22.688 P577timePrecision 13,810,142 22.688 P162 13,810,142 22.688

Top Wikidata items Item Query count % Query count in human subgraph P569 24,521,735 40.286 P570 24,463,540 40.191 P27 18,887,725 31.03 P18 16,424,542 26.984 P50 15,795,288 25.95 P856 15,470,890 25.417 P2002 15,288,837 25.118 P345 15,203,948 24.978 P625 15,191,078 24.957 P2013 15,127,905 24.853 P2003 15,045,926 24.719 P2397 14,983,022 24.615 P580 14,978,083 24.607 P582 14,925,914 24.522 P17 14,875,547 24.439 P136 14,736,247 24.21 P169 14,719,164 24.182 P577 14,672,391 24.105 P212 14,624,020 24.026 P112 14,542,333 23.891

Paths

Paths are more complex predicates that chain properties with logic. Complex paths can increase the scope of a query and also increase its runtime. The table below lists the most used paths in human subgraph queries. While most path are not very complex or long, there are a lot of variety in ways paths are formed to perform queries. Ordinary properties are not considered as paths. The following list contains not only the paths, but also their breakdown into components paths (as done by Jena ARQ while parsing SPARQL queries). For instance: (p:P31/ps:P31)/(wdt:P279)* is recorded as:

• (p:P31/ps:P31)/(wdt:P279)*
  • (p:P31/ps:P31)
    • p:P31
    • ps:P31
  • (wdt:P279)*
    • wdt:P279

The unit form, wdt:P279 for example, was removed from the path list since they are part of other paths and not paths themselves. More paths that seemed obvious as being part of a longer path, and not paths themselves, were also removed from the list for better visualization of the distinct paths used in the queries.

Top Paths

Path	Query count	% Query count in human subgraph
p:P570/psv:P570	13,867,481	22.783
p:P569/psv:P569	13,863,357	22.77
p:P625/psv:P625	13,810,408	22.689
p:P577/psv:P577	13,810,371	22.689
p:P571/psv:P571	13,810,310	22.689
p:P576/psv:P576	13,810,242	22.689
p:P582/psv:P582	13,810,146	22.688
psv:P2046/<http://wikiba.se/ontology#quantityUnit	13,810,142	22.688
p:P580/psv:P580	13,810,142	22.688
psv:P281/<http://wikiba.se/ontology#quantityUnit	13,810,142	22.688
p:P619/psv:P619	13,810,138	22.688
p:P620/psv:P620	13,810,138	22.688
wdt:P31/(wdt:P279)*	1,148,817	1.887
wdt:P31|wdt:P279	1,040,140	1.709
p:P169|p:P488	704,491	1.157
ps:P169|ps:P488	704,491	1.157
p:P2572/ps:P2572	501,987	0.825
((((((((((((wdt:P17|wdt:P101)|wdt:P112)|wdt:P135)|wdt:P136)|wdt:P279)|wdt:P361)|wdt:P460)|wdt:P793)|wdt:P800)|wdt:P1269)|wdt:P1344)|wdt:P1830)|(p:P2572/ps:P2572)	501,987	0.825
ps:P106/(wdt:P279)*	429,856	0.706
ps:P31/(wdt:P279)*	429,284	0.705
wdt:P106/(wdt:P279)*	251,007	0.412
p:P569/ps:P569 p:P570/ps:P570	246,947	0.406
p:P569/ps:P569 p:P570/ps:P570	202,546	0.333
wdt:P50|wdt:P2093	197,426	0.324

Taxon subgraph (Q16521) query analysis

The following analysis was done with query data of November, 2021.

The queries that were estimated to be related to the taxon subgraph accounted for 14.26% of all queries in Wikidata. 13.57% queries used only the taxon subgraph and the rest 0.69% queries used a mix of taxon and various other subgraphs. As described in #What are subgraph related queries, subgraphs are related to queries through Properties, Subject or Object URIs, Subgraph instance items, etc. Here is a breakdown for taxon subgraph taken from #Query count and time. A query can be said to be related to taxon subgraph due to multiple of the following reasons.

• Number of queries: 27,172,995 (14.26%)
• Percent of queries matching subgraph Q-ID, i.e, has Q5: 12.19%
• Percent of queries matching instance items: 0.75%
• Percent of queries matching subject/object URIs: 12.86%
• Percent of queries matching properties: 0.87%
• Percent of queries matching literal strings: 0.43%

Percent of queries matching subject/object URIs (12.86) includes Q-ID (12.19) and instance items (0.75) in them. This makes Q-ID match almost the only reason for queries to be in taxon subgraph. Therefore we look at the top URIs that cause these matches. Also looking at the distribution of such items' usage in queries shows how narrow or wide the search space is (in this case, quite narrow as almost all queries match the Q-ID itself).

Here is a detailed breakdown of what kind of match caused a query to be part of the taxon subgraph:

Taxon subgraph query breakdown item predicate URI taxon Q-id literal # query % all query % taxon query 0 0 1 1 0 22,955,853 12.046 84.48 0 1 0 0 0 1,415,482 0.743 5.209 1 0 0 0 0 638,533 0.335 2.35 0 0 1 0 0 624,147 0.328 2.297 1 0 1 0 0 501,232 0.263 1.845 0 0 0 0 1 443,593 0.233 1.632 0 0 1 1 1 233,109 0.122 0.858 1 1 1 0 0 132,462 0.07 0.487 1 0 0 0 1 66,408 0.035 0.244 0 1 0 0 1 60,111 0.032 0.221 1 1 0 0 0 38,147 0.02 0.14 1 0 1 1 0 30,652 0.016 0.113 0 0 1 0 1 13,104 0.007 0.048 1 0 1 0 1 9,026 0.005 0.033 0 1 1 1 0 5,847 0.003 0.022 1 1 1 1 0 5,248 0.003 0.019 0 1 1 1 1 24 0.0 0.0 0 0 0 1 0 14 0.0 0.0 0 1 1 0 0 3 0.0 0.0 Total 27,172,995 14.26 100

Instance items matched

• Total items used: 588,668
• Total queries that use these items: 1,421,708 (0.75% of all queries)
• The distribution shows there are only 3 high usage(>100k queries) items, and the rest form a long tail of small usage (<1k queries) items in the taxon subgraph.
• Note that these are for the queries from the month of November 2021. These data change from one month to another.

Top items that cause a query to be related to Taxon subgraph (Q16521) Instance item Instance item label #of queries Q15978631 Homo sapiens 148,455 Q83310 house mouse 111,473 Q184224 brown rat 111,397 Q25400 Asteraceae 67,794 Q729 animal 43,360 Q25308 Orchidaceae 22,905 Q756 plant 11,244 Q14560 Cactaceae 7,840 Q173756 Apocynaceae 7,018 Q526228 Acantharchus pomotis 4,373 Q36341 Brown Bear 1,878 Q80174 Pan 1,696 Q19537 bonobo 1,684 Q69581 Siberian tiger 1,664 Q171497 sunflower 1,637 Q504549 Spur-thighed tortoise 1,617 Q8202634 Apis mellifera sahariensis 1,438 Q41960 Ailurus fulgens 1,317 Q2346039 Thunnus 1,285 Q719725 Saccharomyces cerevisiae 1,244

Properties matched

• Total properties used: 162 (Recall these are properties that occur 99% of the times in the taxon subgraph)
• Total queries that use these properties: 1,657,324 (0.87% of all queries)
• Most of these look like external IDs. Only 31 of these properties are not IDs.
• The distribution shows there is 1 property with >1M queries, 7 properties with >100K queries, 14 properties with 2-8K queries, and rest of the properties match ~1K and less queries. In short, the distribution is a extremely skewed by only ~10 properties.

Top properties that cause a query to be related to Taxon subgraph (Q16521) Property Property label #of queries P3151 iNaturalist taxon ID 1,346,905 P141 IUCN conservation status 167,512 P183 endemic to 152,618 P961 IPNI plant ID 64,340 P938 FishBase species ID 39,494 P6018 SeaLifeBase ID 25,217 P574 year of taxon publication 12,599 P2040 CITES Species+ ID 12,073 P697 ex taxon author 10,710 P566 basionym 8,469 P5473 The Reptile Database ID 6,357 P5036 AmphibiaWeb Species ID 6,343 P7715 World Flora Online ID 4,868 P5626 Global Invasive Species Database ID 4,509 P5037 Plants of the World online ID 3,629 P6105 Observation.org ID 3,193 P960 Tropicos ID 2,922 P9157 Open Tree of Life ID 2,470 P1070 PlantList-ID 2,181 P1772 USDA PLANTS ID 2,167

Subject/Object URI matched

• Total URIs used: 651,945 (Recall these are URIs that occur 99% of the times in the human subgraph)
• Total queries that use these URIs: 24,510,707 (12.86% of all queries)
• The top URI is in fact the Q-ID of taxon subgraph - Q16521 - and matches 12.19%of all queries. We look into the queries directly later in this section.
• We analyze the top 100K URIs. Of these, 66% are Wikidata items, 31% are Wikipedia links.
• The distribution shows that the top 2 URIs occur in queries tens of times greater than the other URIs. Of course this data is only for November 2021, but the high usage of Taxon Q-ID was also observed in October 2021 data.

Top URIs that cause a query to be related to Taxon subgraph (Q16521) URI URI label #of queries Q16521 taxon 23,230,733 (12.86%) Q767728 variety 50,370 (0.03%) Q279749 form 7,100 https://ja.wikipedia.org/wiki/キーウィ_(鳥) - 4,827 https://en.wikipedia.org/wiki/Tokay_gecko - 2,602 https://ja.wikipedia.org/wiki/ガラパゴスリクイグアナ - 2,309 https://ja.wikipedia.org/wiki/セミ - 2,186 https://ja.wikipedia.org/wiki/ドードー - 1,932 https://ja.wikipedia.org/wiki/ダイオウホウズキイカ - 1,864 https://ja.wikipedia.org/wiki/アオリイカ - 1,843 https://ja.wikipedia.org/wiki/コウイカ目 - 1,835 https://en.wikipedia.org/wiki/Bobtail_squid - 1,834 https://ja.wikipedia.org/wiki/緩歩動物 - 1,821 https://ja.wikipedia.org/wiki/トグロコウイカ - 1,819 https://ja.wikipedia.org/wiki/アメリカオオアカイカ - 1,814 https://ja.wikipedia.org/wiki/ケンサキイカ - 1,812 https://ja.wikipedia.org/wiki/ハモ - 1,547 https://en.wikipedia.org/wiki/Coronavirus - 1,421 https://ja.wikipedia.org/wiki/ダンゴムシ - 1,402 https://ja.wikipedia.org/wiki/ブタ - 1,333

The Wikipedia links are from 80 different languages. The tables below show some of the top languages used in terms of unique links queried and shows the top 5 links for each of these languages.

Top language Wikipedias as URIs in taxon subgraph queries

Language: en, # unique links:13533 URI # Query en.wikipedia.org/wiki/Tokay_gecko 2,602 en.wikipedia.org/wiki/Bobtail_squid 1,834 en.wikipedia.org/wiki/Coronavirus 1,421 en.wikipedia.org/wiki/Fusarium 1,204 en.wikipedia.org/wiki/Tardigrade 1,202	Language: ja, # unique links:6611 URI # Query ja.wikipedia.org/wiki/キーウィ_(鳥) 4,827 ja.wikipedia.org/wiki/ガラパゴスリクイグアナ 2,309 ja.wikipedia.org/wiki/セミ 2,186 ja.wikipedia.org/wiki/ドードー 1,932 ja.wikipedia.org/wiki/ダイオウホウズキイカ 1,864	Language: es, # unique links:2908 URI # Query es.wikipedia.org/wiki/Eudocimus_ruber 977 es.wikipedia.org/wiki/Chelonioidea 610 es.wikipedia.org/wiki/Pelecanus 458 es.wikipedia.org/wiki/Fregata_magnificens 262 es.wikipedia.org/wiki/Ebolavirus 255	Language: de, # unique links:1950 URI # Query de.wikipedia.org/wiki/Coronaviridae 123 de.wikipedia.org/wiki/Taubenschwänzchen 95 de.wikipedia.org/wiki/Mariendistel 79 de.wikipedia.org/wiki/Wespenspinne 75 de.wikipedia.org/wiki/Schnaken 72
Language: fr, # unique links:1342 URI # Query fr.wikipedia.org/wiki/Homo_sapiens 201 fr.wikipedia.org/wiki/Coronavirus 107 fr.wikipedia.org/wiki/Tardigrada 80 fr.wikipedia.org/wiki/Scutigère_véloce 76 fr.wikipedia.org/wiki/Muguet_de_mai 55	Language: nl, # unique links:1056 URI # Query nl.wikipedia.org/wiki/Europese_hoornaar 105 nl.wikipedia.org/wiki/Vuurwants 66 nl.wikipedia.org/wiki/Stinkende_kortschildkever 61 nl.wikipedia.org/wiki/Stadsreus_(zweefvlieg) 48 nl.wikipedia.org/wiki/Coronavirussen 47	Language: ru, # unique links:800 URI # Query ru.wikipedia.org/wiki/Мимивирус 53 ru.wikipedia.org/wiki/Коронавирусы 50 ru.wikipedia.org/wiki/Обыкновенная_мухоловка 49 ru.wikipedia.org/wiki/Тихоходки 40 ru.wikipedia.org/wiki/Малая_панда 40	Language: pt, # unique links:520 URI # Query pt.wikipedia.org/wiki/Candiru 39 pt.wikipedia.org/wiki/Tangerina 36 pt.wikipedia.org/wiki/Tardigrada 34 pt.wikipedia.org/wiki/Grelo 33 pt.wikipedia.org/wiki/Panda-vermelho 31

Query time

• The total query time of taxon subgraph is ~3% of total query time and total query count is 14.26% of all queries.
• Average time per query is 0.064 seconds (64 ms). Almost all queries in this subgraph are small and simple.
• The query time distribution is shown in the chart below: in absolute counts, in percent of queries in taxon subgraph, and in percent of all queries.

User agent

List of top user agents that query the taxon subgraph is given below. This helps us view the distribution of usage - whether few user agents dominate the usage or it is a rather well distributed usage scenario across user agents. Top 10 user agents in terms of query count and query time is shown in the table below. The query type column is discussed later in the section.

Top user agents in taxon subgraph

User agent	Query count	% query in taxon subgraph	% query overall	Query time(hr)	% query time in taxon subgraph	% query time overall	# query type
mix-n-match	22,959,293	84.493	12.048	163	33.949	1.063	5
Hub	1,318,251	4.851	0.692	17	3.455	0.108	1
WikidataQueryServiceR	568,799	2.093	0.298	9	1.837	0.058	34
UA#4	325,563	1.198	0.171	10	2.044	0.064	5
UA#5	265,565	0.977	0.139	2	0.495	0.015	5
UA#6	199,650	0.735	0.105	53	11.133	0.349	24
UA#7	168,536	0.62	0.088	2	0.441	0.014	2
sparqlwrapper	161,781	0.595	0.085	126	26.257	0.822	33
UA#9	107,736	0.396	0.057	3	0.627	0.02	1
EasyContent	103,292	0.38	0.054	2	0.346	0.011	1
UA#11	1,330	0.005	0.001	12	2.481	0.078	8
UA#12	45,065	0.166	0.024	8	1.644	0.051	9
AhrefsBot	56,580	0.208	0.03	8	1.575	0.049	12
Apache-Jena-ARQ	6,654	0.024	0.003	6	1.27	0.04	11

The query time breakdown was plotted for the top 20 user agents (in terms of time).

Query types

Query types are grouped by the operations a query uses and also the order of operations used. This groups similar queries together despite different information sought and also separates groups of simple or complicated queries. The taxon subgraph has ~1100 different types of queries (The variety is quite less compared to 11K query type in human subgraph). Notice that some query groups can be very similar in what they ask for as well, although most groups differ a lot. The top query groups are listed below. Only the top 3 types of queries account for ~90% of the queries of taxon subgraph. Top 40 form 99% of the queries in this subgraph. The rest form a long tail of small query counts.

Top query types in taxon subgraph

Query group (operation list)	#of queries	%of queries in human subgraph
['path', 'table', 'bgp', 'join', 'bgp', 'union', 'join', 'project']	13,013,162	47.89
['path', 'table', 'bgp', 'join', 'bgp', 'union', 'bgp', 'union', 'join', 'project']	9,943,644	36.594
['bgp', 'project', 'distinct', 'slice']	1,318,254	4.851
['table', 'bgp', 'leftjoin', 'bgp', 'join', 'filter', 'project']	236,468	0.87
['bgp', 'project']	230,329	0.848
['path', 'bgp', 'path', 'sequence', 'table', 'join', 'filter', 'order', 'project', 'distinct']	199,441	0.734
['bgp', 'bgp', 'service', 'join', 'project', 'slice']	169,045	0.622
['table', 'extend', 'extend', 'bgp', 'join', 'project']	152,885	0.563
['table', 'path', 'bgp', 'sequence', 'path', 'bgp', 'sequence', 'path', 'bgp', 'sequence', 'leftjoin', 'leftjoin', 'leftjoin', 'bgp', 'leftjoin', 'bgp', 'leftjoin', 'bgp', 'leftjoin', 'bgp', 'leftjoin', 'bgp', 'leftjoin', 'bgp', 'service', 'join', 'project', 'distinct']	152,353	0.561
['bgp', 'extend', 'bgp', 'extend', 'union', 'bgp', 'extend', 'union', 'bgp', 'extend', 'union', 'path', 'extend', 'union', 'path', 'extend', 'union', 'bgp', 'extend', 'union', 'bgp', 'extend', 'union', 'bgp', 'extend', 'union', 'bgp', 'extend', 'union', 'bgp', 'extend', 'union', 'bgp', 'extend', 'union', 'bgp', 'extend', 'union', 'bgp', 'extend', 'union', 'bgp', 'extend', 'union', 'path', 'extend', 'union', 'bgp', 'extend', 'union', 'path', 'extend', 'union', 'bgp', 'extend', 'union', 'path', 'extend', 'union', 'project', 'distinct']	139,177	0.512

Looking at the top 5 query groups (that make up >90% of all taxon subgraph queries), the following query types were found:

1. Search with taxon name, synonyms, altlabels etc
2. Search with taxon ID. E.g. SELECT DISTINCT ?subject WHERE { ?subject wdt:P3151 '47126' .}
3. Get labels of certain items
4. Get labels of certain items in specific languages
5. Get external IDs of items

UA vs query types

Getting the number of query types per user agent informs us of the variety of queries a user agent makes to WDQS. This also breaks down the taxon subgraph queries into finer groups. The following plot shows the number of query types for each user agent in the taxon subgraph.

This shows us that most user agents make only 1 type of query. Only 3 user agents make queries of >100 types, and 5 user agents make queries of 50-100 types. The figure below shows the number of query per query type for the top 8 user agents.

** The number of query types for the top user agents in terms of query count and time is listed in the Taxon User agent section.

Query type vs time class

While there are close to 1,100 query types and only the top 3 types of queries account for ~90% of the queries of taxon subgraph (12.8% of all queries), not all of them are equally time consuming. Some can be simple queries, while some can be long and complex. The following plot shows the top 10 query types with query time classes. The values above the bar show both percent in taxon subgraph and overall query percentage. The subplots are titled with percent of the number of queries in that query type.

In sum, most queries in taxon subgraph are small and simple, take between 10 to 100ms time to run, and are mostly by 1/2 user agents.

Services

The queries use 12 unique services. The top 4 services are the most used, although the usage is still pretty low; rest are used in less than 30 queries.

Services used in queries in taxon subgraph

Service	Query count	% query in taxon subgraph
wikibase:label	853,272	3.14
wikibase:mwapi	8,198	0.03
gas:service	1,206	0.004
bd:sample	512	0.002
https://query.wikidata.org/bigdata/namespace/wdq/sparql	28	0.0
idsm:wikidata	28	0.0
https://query.wikidata.org/sparql	14	0.0
http://sparql.wikipathways.org/sparql	13	0.0
wikibase:around	3	0.0
https://sparql.wikipathways.org/sparql	2	0.0
https://sophox.org/sparql	2	0.0
https://spang.dbcls.jp/sparql	1	0.0

Triples

Some query type analysis done in section query types gives us a good idea of what kind of queries taxon subgraph receives. Looking at the triples themselves also helps us peek into what most of the queries look like, what the most common subjects, objects, and properties are. The table below lists these along with the top wikidata items and properties used overall.

Top triples in taxon subgraph queries

Subject	Predicate	Object	# query	% query of taxon subgraph
q	wdt:P31/(wdt:P279)*	wd:Q16521	22,956,806	84.484*
bd:serviceParam	wikibase:language	en	379,833	1.398
item	wdt:P225	taxonName	265,426	0.977
item	p:P105	taxonRank1	265,360	0.977
taxonRank1	ps:P105	taxonRank	265,360	0.977
item	rdfs:label	label	241,505	0.889
item	skos:altLabel	altLabel	236,495	0.870
bd:serviceParam	wikibase:language	[AUTO_LANGUAGE],en	205,404	0.756
items	rdfs:label	itemlabel	199,441	0.734
items	(wdt:P279)?	types	195,916	0.721

*coincides with the number of queries from mix-n-match. Almost all mix-in-match queries in taxon subgraph have this triple. And all of these triples only occur in mix-n-match queries.

Top Subjects Subject Query count % Query in taxon subgraph q 23,071,376 84.906 subject 1,335,606 4.915 item 965,952 3.555 bd:serviceParam 853,712 3.142 taxonRank1 265,360 0.977 items 199,444 0.734 resource 153,967 0.567 class 153,628 0.565 order 152,354 0.561 family 152,354 0.561 P27 112,441 0.414 wikidata_id 111,123 0.409 https://ja.wikipedia.org/wiki/ 106,127 0.391 wdpage 71,631 0.264 id 39,642 0.146 X 29,097 0.107 Z 28,987 0.107 supertype 23,178 0.085 property 21,799 0.08 location 17,093 0.063

Top Predicates Predicate Query count % Query in taxon subgraph skos:altLabel 23,214,129 85.431 wdt:P31/(wdt:P279)* 23,033,714 84.767 prop 22,957,095 84.485 wdt:P3151 1,342,739 4.941 rdfs:label 881,092 3.243 wikibase:language 853,273 3.14 wdt:P225 751,916 2.767 wdt:P31 643,819 2.369 schema:about 611,190 2.249 wdt:P1843 430,635 1.585 wdt:P18 322,440 1.187 p:P105 280,663 1.033 ps:P105 280,547 1.032 wdt:P105 207,658 0.764 (wdt:P279)? 195,916 0.721 ((wdt:P31)*/(wdt:P279)*)/(wd:P361)* 195,915 0.721 (wdt:P171)+ 168,128 0.619 wdt:P141 165,257 0.608 wdt:P183 152,479 0.561 wdt:P1889 152,376 0.561

Top Objects Object Query count % Query in taxon subgraph wd:Q16521 23,000,295 84.644 taxonName 422,507 1.555 en 387,760 1.427 pic 290,568 1.069 label 277,361 1.021 taxonRank 270,223 0.994 taxonRank1 265,360 0.977 altLabel 237,031 0.872 class 218,616 0.805 [AUTO_LANGUAGE],en 205,404 0.756 itemlabel 199,450 0.734 items 199,441 0.734 types 195,920 0.721 name 181,578 0.668 wd:Q35409 162,940 0.6 wd:Q36602 155,742 0.573 wd:Q37517 153,105 0.563 endemicTo 152,386 0.561 conservationStatus 152,357 0.561 order 152,354 0.561

Top Wikidata items Item Query count % Query in human subgraph P31 23,985,585 39.406 P225 23,735,291 38.994 P279 23,529,502 38.656 Q16521 23,230,747 38.165 P1420 22,965,572 37.73 P3151 1,346,905 2.213 P105 490,972 0.807 P1843 433,159 0.712 P18 325,372 0.535 P171 311,272 0.511 P625 273,859 0.45 P361 212,381 0.349 Q379813 200,195 0.329 Q152 199,784 0.328 Q2095 199,645 0.328 Q25403900 199,589 0.328 Q11004 199,561 0.328 Q12117 199,544 0.328 Q10990 199,527 0.328 Q3314483 199,503 0.328

Paths

Paths are more complex predicates that chain properties with logic. Complex paths can increase the scope of a query and also increase its runtime. The table below lists the most used paths in taxon subgraph queries. While most path are not very complex or long, there are a lot of variety in ways paths are formed to perform queries. Ordinary properties are not considered as paths. The following list contains not only the paths, but also their breakdown into components paths (as done by Jena ARQ while parsing SPARQL queries). For instance: (p:P31/ps:P31)/(wdt:P279)* is recorded as:

• (p:P31/ps:P31)/(wdt:P279)*
  • (p:P31/ps:P31)
    • p:P31
    • ps:P31
  • (wdt:P279)*
    • wdt:P279

The unit form, wdt:P279 for example, was removed from the path list since they are part of other paths and not paths themselves. More paths that seemed obvious as being part of a longer path, and not paths themselves, were also removed from the list for better visualization of the distinct paths used in the queries.

Top Paths

Path	Query count	% Query in taxon subgraph
wdt:P31/(wdt:P279)*	23,035,627	84.774
((wdt:P31)*/(wdt:P279)*)/(wd:P361)*	195,916	0.721
(wdt:P171)+	168,128	0.619
wdt:P1416|wdt:P108	139,191	0.512
(wdt:P159)?/wdt:P625	139,191	0.512
wdt:P50|wdt:P2093	139,191	0.512
^wdt:P31/wdt:P235	139,191	0.512
wdt:P31/(wdt:P279)?	139,184	0.512
(((((((((((((((((((((((((wdt:P171|(wdt:P171/wdt:P171))|((wdt:P171/wdt:P171)/wdt:P171))|(((wdt:P171/wdt:P171)/wdt:P171)/wdt:P171))|(wdt:P171/wdt:P171*4))|(wdt:P171/wdt:P171*5))|(wdt:P171/wdt:P171*6))|(wdt:P171/wdt:P171*7))|(wdt:P171/wdt:P171*8))|(wdt:P171/wdt:P171*9))|(wdt:P171/wdt:P171*10))|(wdt:P171/wdt:P171*11))|(wdt:P171/wdt:P171*12))|(wdt:P171/wdt:P171*13))|(wdt:P171/wdt:P171*14))|(wdt:P171/wdt:P171*15))|(wdt:P171/wdt:P171*16))|(wdt:P171/wdt:P171*17))|(wdt:P171/wdt:P171*18))|(wdt:P171/wdt:P171*19))|(wdt:P171/wdt:P171*20))|(wdt:P171/wdt:P171*21))|(wdt:P171/wdt:P171*22))|(wdt:P171/wdt:P171*23))|(wdt:P171/wdt:P171*24))|(wdt:P171/wdt:P171*25))	112,475	0.414
wdt:P31|wdt:P279	103,573	0.381
(wdt:P1647)*	11,178	0.041
(((((((((((((((wdt:P17|wdt:P101)|wdt:P112)|wdt:P135)|wdt:P136)|wdt:P279)|wdt:P361)|wdt:P460)|wdt:P793)|wdt:P800)|wdt:P1269)|wdt:P1344)|wdt:P1830)|(p:P2572/ps:P2572))|wdt:P3342)|wdt:P3602)|wdt:P5004	11,002	0.04