A network of correspondence: the Datini letters

Francesco di Marco Datini, a merchant of Prato who died in 1410, ran a firm with offices across the western Mediterranean, and kept its letters. The Datini Correspondence Metadata describes 145,417 of them, one record each: which city a letter was sent from and to, the day it was sent and the day it arrived. Counted city by city, they make a network: places joined by the post that ran between them.

Unlike the River Idle, where a reach of river is a thing in the landscape, the connection between Florence and Barcelona is a relationship, not a road. There is nothing between the two cities to record as a place of its own, so the connection is recorded directly, and what the source says about it goes with it.

Download the example as CSV files. The same data is in the repository as PLATO JSON, and part of it, with comments, as linked data.

The sources

source_id	title	citation	uri	date	from	to	derived_from	licence
datini-metadata	Datini Correspondence Metadata	A. Franklin-Lyons, Datini Correspondence Metadata Project (BodoArXiv, 2024), doi:10.34055/osf.io/wgpm4, with the data at doi:10.17605/OSF.IO/MT25A: one record for each letter of the Datini archive, from the catalogue of the Archivio di Stato di Prato	https://doi.org/10.17605/OSF.IO/MT25A	2024	2024	2024		https://creativecommons.org/licenses/by/4.0/
letters-by-pair	Letters and delivery times by pair of cities	Counts of letters, and median delivery times, for pairs of cities, worked out for this example on 29 September 2026 from the Datini Correspondence Metadata		2026	2026	2026	datini-metadata	https://creativecommons.org/licenses/by/4.0/

The dataset is the first source. The second is this example’s own work: the counts of letters and the delivery times, worked out from the dataset’s records on a stated day. It is a source of its own, derived_from the dataset, because the figures are findings someone made from the records, not statements the records make. Anyone who wants to check them can see what they came from and how.

The network and its cities

place_id	label	country_codes
datini-network	The Datini correspondence network: the principal offices	IT;ES;FR
florence	Florence	IT
pisa	Pisa	IT
prato	Prato	IT
genoa	Genoa	IT
barcelona	Barcelona	ES
valencia	Valencia	ES
palma	Palma de Mallorca	ES
avignon	Avignon	FR

place_id	type_label	type_uri	date	from	to	source_id	citation_function	notes
datini-network	network	https://w3id.org/plato#TypeNetwork	1355-1443	1355	1443	datini-metadata	citesAsDataSource	The span of sending dates the dataset gives for the letters counted here.

place_id	relation_type	related_place_id	date	source_id	citation_function	notes
florence	MemberOf	datini-network	undated	datini-metadata	citesAsDataSource	Nothing in the letters puts the cities in an order, so no sequence.
pisa	MemberOf	datini-network	undated	datini-metadata	citesAsDataSource	Nothing in the letters puts the cities in an order, so no sequence.
prato	MemberOf	datini-network	undated	datini-metadata	citesAsDataSource	Nothing in the letters puts the cities in an order, so no sequence.
genoa	MemberOf	datini-network	undated	datini-metadata	citesAsDataSource	Nothing in the letters puts the cities in an order, so no sequence.
barcelona	MemberOf	datini-network	undated	datini-metadata	citesAsDataSource	Nothing in the letters puts the cities in an order, so no sequence.
valencia	MemberOf	datini-network	undated	datini-metadata	citesAsDataSource	Nothing in the letters puts the cities in an order, so no sequence.
palma	MemberOf	datini-network	undated	datini-metadata	citesAsDataSource	Nothing in the letters puts the cities in an order, so no sequence.
avignon	MemberOf	datini-network	undated	datini-metadata	citesAsDataSource	Nothing in the letters puts the cities in an order, so no sequence.

Each city is a MemberOf the network, with no sequence: nothing in the letters puts the cities in an order.

The connections

place_id	relation_type	related_place_id	property_uri	property_label	value	unit_uri	date	from	to	source_id	locator	citation_function	notes
florence	LeadsTo	pisa	https://www.wikidata.org/wiki/Property:P1114	letters sent	16261		1355-1402	1355	1402	letters-by-pair	FIRENZE to PISA	citesAsDataSource	Every letter the dataset records from FIRENZE to PISA. The dates are the years of sending as the dataset gives them.
florence	LeadsTo	pisa	https://www.wikidata.org/wiki/Property:P2047	median delivery time (days)	2	http://www.wikidata.org/entity/Q573	1355-1402	1355	1402	letters-by-pair	FIRENZE to PISA	citesAsDataSource	The median of arrival date minus sending date, over the 13555 letters with both dates and a delay of 0 to 365 days. The mean is not given: impossible delays in the records move it too far.
pisa	LeadsTo	florence	https://www.wikidata.org/wiki/Property:P1114	letters sent	6414		1379-1413	1379	1413	letters-by-pair	PISA to FIRENZE	citesAsDataSource	Every letter the dataset records from PISA to FIRENZE. The dates are the years of sending as the dataset gives them.
pisa	LeadsTo	florence	https://www.wikidata.org/wiki/Property:P2047	median delivery time (days)	2	http://www.wikidata.org/entity/Q573	1379-1413	1379	1413	letters-by-pair	PISA to FIRENZE	citesAsDataSource	The median of arrival date minus sending date, over the 4933 letters with both dates and a delay of 0 to 365 days. The mean is not given: impossible delays in the records move it too far.
florence	LeadsTo	prato	https://www.wikidata.org/wiki/Property:P1114	letters sent	6804		1367-1443	1367	1443	letters-by-pair	FIRENZE to PRATO	citesAsDataSource	Every letter the dataset records from FIRENZE to PRATO. The dates are the years of sending as the dataset gives them.
florence	LeadsTo	prato	https://www.wikidata.org/wiki/Property:P2047	median delivery time (days)	0	http://www.wikidata.org/entity/Q573	1367-1443	1367	1443	letters-by-pair	FIRENZE to PRATO	citesAsDataSource	The median of arrival date minus sending date, over the 5703 letters with both dates and a delay of 0 to 365 days. The mean is not given: impossible delays in the records move it too far.
prato	LeadsTo	florence	https://www.wikidata.org/wiki/Property:P1114	letters sent	5700		1383-1419	1383	1419	letters-by-pair	PRATO to FIRENZE	citesAsDataSource	Every letter the dataset records from PRATO to FIRENZE. The dates are the years of sending as the dataset gives them.
prato	LeadsTo	florence	https://www.wikidata.org/wiki/Property:P2047	median delivery time (days)	1	http://www.wikidata.org/entity/Q573	1383-1419	1383	1419	letters-by-pair	PRATO to FIRENZE	citesAsDataSource	The median of arrival date minus sending date, over the 4848 letters with both dates and a delay of 0 to 365 days. The mean is not given: impossible delays in the records move it too far.
florence	LeadsTo	genoa	https://www.wikidata.org/wiki/Property:P1114	letters sent	5306		1379-1405	1379	1405	letters-by-pair	FIRENZE to GENOVA	citesAsDataSource	Every letter the dataset records from FIRENZE to GENOVA. The dates are the years of sending as the dataset gives them.
florence	LeadsTo	genoa	https://www.wikidata.org/wiki/Property:P2047	median delivery time (days)	8	http://www.wikidata.org/entity/Q573	1379-1405	1379	1405	letters-by-pair	FIRENZE to GENOVA	citesAsDataSource	The median of arrival date minus sending date, over the 4578 letters with both dates and a delay of 0 to 365 days. The mean is not given: impossible delays in the records move it too far.
genoa	LeadsTo	florence	https://www.wikidata.org/wiki/Property:P1114	letters sent	2902		1382-1414	1382	1414	letters-by-pair	GENOVA to FIRENZE	citesAsDataSource	Every letter the dataset records from GENOVA to FIRENZE. The dates are the years of sending as the dataset gives them.
genoa	LeadsTo	florence	https://www.wikidata.org/wiki/Property:P2047	median delivery time (days)	7	http://www.wikidata.org/entity/Q573	1382-1414	1382	1414	letters-by-pair	GENOVA to FIRENZE	citesAsDataSource	The median of arrival date minus sending date, over the 2427 letters with both dates and a delay of 0 to 365 days. The mean is not given: impossible delays in the records move it too far.
florence	LeadsTo	barcelona	https://www.wikidata.org/wiki/Property:P1114	letters sent	5097		1382-1412	1382	1412	letters-by-pair	FIRENZE to BARCELLONA	citesAsDataSource	Every letter the dataset records from FIRENZE to BARCELLONA. The dates are the years of sending as the dataset gives them.
florence	LeadsTo	barcelona	https://www.wikidata.org/wiki/Property:P2047	median delivery time (days)	30	http://www.wikidata.org/entity/Q573	1382-1412	1382	1412	letters-by-pair	FIRENZE to BARCELLONA	citesAsDataSource	The median of arrival date minus sending date, over the 4314 letters with both dates and a delay of 0 to 365 days. The mean is not given: impossible delays in the records move it too far.
barcelona	LeadsTo	valencia	https://www.wikidata.org/wiki/Property:P1114	letters sent	3551		1386-1419	1386	1419	letters-by-pair	BARCELLONA to VALENZA	citesAsDataSource	Every letter the dataset records from BARCELLONA to VALENZA. The dates are the years of sending as the dataset gives them.
barcelona	LeadsTo	valencia	https://www.wikidata.org/wiki/Property:P2047	median delivery time (days)	7	http://www.wikidata.org/entity/Q573	1386-1419	1386	1419	letters-by-pair	BARCELLONA to VALENZA	citesAsDataSource	The median of arrival date minus sending date, over the 2882 letters with both dates and a delay of 0 to 365 days. The mean is not given: impossible delays in the records move it too far.
valencia	LeadsTo	barcelona	https://www.wikidata.org/wiki/Property:P1114	letters sent	2841		1382-1411	1382	1411	letters-by-pair	VALENZA to BARCELLONA	citesAsDataSource	Every letter the dataset records from VALENZA to BARCELLONA. The dates are the years of sending as the dataset gives them.
valencia	LeadsTo	barcelona	https://www.wikidata.org/wiki/Property:P2047	median delivery time (days)	7	http://www.wikidata.org/entity/Q573	1382-1411	1382	1411	letters-by-pair	VALENZA to BARCELLONA	citesAsDataSource	The median of arrival date minus sending date, over the 2211 letters with both dates and a delay of 0 to 365 days. The mean is not given: impossible delays in the records move it too far.
barcelona	LeadsTo	palma	https://www.wikidata.org/wiki/Property:P1114	letters sent	3193		1394-1411	1394	1411	letters-by-pair	BARCELLONA to MAIORCA	citesAsDataSource	Every letter the dataset records from BARCELLONA to MAIORCA. The dates are the years of sending as the dataset gives them.
barcelona	LeadsTo	palma	https://www.wikidata.org/wiki/Property:P2047	median delivery time (days)	5	http://www.wikidata.org/entity/Q573	1394-1411	1394	1411	letters-by-pair	BARCELLONA to MAIORCA	citesAsDataSource	The median of arrival date minus sending date, over the 2947 letters with both dates and a delay of 0 to 365 days. The mean is not given: impossible delays in the records move it too far.
avignon	LeadsTo	florence	https://www.wikidata.org/wiki/Property:P1114	letters sent	2517		1371-1420	1371	1420	letters-by-pair	AVIGNONE to FIRENZE	citesAsDataSource	Every letter the dataset records from AVIGNONE to FIRENZE. The dates are the years of sending as the dataset gives them.
avignon	LeadsTo	florence	https://www.wikidata.org/wiki/Property:P2047	median delivery time (days)	20	http://www.wikidata.org/entity/Q573	1371-1420	1371	1420	letters-by-pair	AVIGNONE to FIRENZE	citesAsDataSource	The median of arrival date minus sending date, over the 2114 letters with both dates and a delay of 0 to 365 days. The mean is not given: impossible delays in the records move it too far.

This is the connections sheet. Each row joins two cities with LeadsTo, from the city in place_id to the one in related_place_id, and gives one figure about that connection: the number of letters, or the median number of days they took to arrive. Two figures about one connection are two rows.

Letters went both ways between Florence and Pisa, but not equally: 16,261 from Florence to Pisa, 6,414 back. Each direction is its own connection, with its own figures, which is why these rows use LeadsTo twice and not ConnectedTo once.

The delivery time is a median: half the letters took less time, half more. The records contain impossible delays, letters apparently received before they were sent or years later, which probably come from how their dates were written. They move an average a long way but a median hardly at all. So the median is given, over the letters with both dates and a delay of up to a year, and the note on each row says exactly which letters that is. date, from and to give the years of sending that the figures cover, as the dataset dates them.

Names, locations and identities

place_id	name	language	script	form_status	date	source_id	locator	citation_function	notes
florence	FIRENZE	it	Latn	Headword	undated	datini-metadata	provenienza, destinazione	citesAsDataSource	The name the archive catalogue files the city under.
pisa	PISA	it	Latn	Headword	undated	datini-metadata	provenienza, destinazione	citesAsDataSource	The name the archive catalogue files the city under.
prato	PRATO	it	Latn	Headword	undated	datini-metadata	provenienza, destinazione	citesAsDataSource	The name the archive catalogue files the city under.
genoa	GENOVA	it	Latn	Headword	undated	datini-metadata	provenienza, destinazione	citesAsDataSource	The name the archive catalogue files the city under.
barcelona	BARCELLONA	it	Latn	Headword	undated	datini-metadata	provenienza, destinazione	citesAsDataSource	The name the archive catalogue files the city under.
valencia	VALENZA	it	Latn	Headword	undated	datini-metadata	provenienza, destinazione	citesAsDataSource	The name the archive catalogue files the city under.
palma	MAIORCA	it	Latn	Headword	undated	datini-metadata	provenienza, destinazione	citesAsDataSource	The name the archive catalogue files the city under.
avignon	AVIGNONE	it	Latn	Headword	undated	datini-metadata	provenienza, destinazione	citesAsDataSource	The name the archive catalogue files the city under.

The archive files each city under an Italian name, in capitals (FIRENZE, BARCELLONA); these are recorded as the dataset’s headwords.

place_id	latitude	longitude	geometry_role	date	source_id	locator	citation_function
florence	43.7695604	11.2558136	RepresentativePoint	undated	datini-metadata	City List with Coordinates, FIRENZE	citesAsDataSource
pisa	43.7228386	10.4016888	RepresentativePoint	undated	datini-metadata	City List with Coordinates, PISA	citesAsDataSource
prato	43.8777049	11.102228	RepresentativePoint	undated	datini-metadata	City List with Coordinates, PRATO	citesAsDataSource
genoa	44.4056499	8.946256	RepresentativePoint	undated	datini-metadata	City List with Coordinates, GENOVA	citesAsDataSource
barcelona	41.3850639	2.1734035	RepresentativePoint	undated	datini-metadata	City List with Coordinates, BARCELLONA	citesAsDataSource
valencia	39.4699075	-0.3762881	RepresentativePoint	undated	datini-metadata	City List with Coordinates, VALENZA	citesAsDataSource
palma	39.570023	2.650706	RepresentativePoint	undated	datini-metadata	City List with Coordinates, MAIORCA	citesAsDataSource
avignon	43.949947	4.8086864	RepresentativePoint	undated	datini-metadata	City List with Coordinates, AVIGNONE	citesAsDataSource

place_id	same_as	match_type	basis
florence	http://www.wikidata.org/entity/Q2044	exactMatch	The city the dataset names, and its English name (Florence)
pisa	http://www.wikidata.org/entity/Q13375	exactMatch	The city the dataset names, and its English name (Pisa)
prato	http://www.wikidata.org/entity/Q13377	exactMatch	The city the dataset names, and its English name (Prato)
genoa	http://www.wikidata.org/entity/Q1449	exactMatch	The city the dataset names, and its English name (Genoa)
barcelona	http://www.wikidata.org/entity/Q1492	exactMatch	The city the dataset names, and its English name (Barcelona)
valencia	http://www.wikidata.org/entity/Q8818	exactMatch	The city the dataset names, and its English name (Valencia)
palma	http://www.wikidata.org/entity/Q8826	exactMatch	The city the dataset names, and its English name (Palma de Majorca)
avignon	http://www.wikidata.org/entity/Q6397	exactMatch	The city the dataset names, and its English name (Avignon)

The coordinates are the dataset’s own, and each city is matched to its Wikidata item.