98/ About samples and numbers in artwork historical past

“There Will Be Numbers” Andrew Piper (1)

Any (digital) artwork historic assortment is made up of samples of an indefinite inhabitants of artists and their artworks. Sampling in artwork historical past is rarely ‘at random’, however is all the time a ‘handy sampling’, with a bias’ since info sources are without end restricted. Subsequently, analysing samples is important in artwork historic analysis.

On this put up, two equal however impartial samples of French artists are examined on the nominal, ordinal and interval stage. They’re extracted from the long-term undertaking of the Iconography of Venus from the Center Ages to Trendy Occasions (2). The evaluation demonstrates how the temporal distributions of the artists are considerably totally different in each samples. 

The evaluation of the samples

 Pattern 1 of 386 artists, and Pattern 2 of 669 artists, are chosen datasets of French artists restricted in alphabetical order by letters A, B, C, D of their surnames. The datasets comprise additionally the given names, the gender, dates and locations of delivery and loss of life or, a minimum of, the 12 months and place of exercise as recognized. Pattern 1, represents 39,5% of the total dataset of 977 artists, compiled in 2004-2008 and revealed within the Topical Catalogue ‘The French Venus’ Vol. 2.1 in 2009 (3). Pattern 2 is representing 40,3% of the provisional dataset of 1658 artists, compiled in 2009-2021 and revealed in ‘The French Venus model 2.1.3’ in April 2021 (4). Thus their relative dimension with respect to the total datasets is equal, although each samples differ largely in dimension. Moreover, their numbers per letter, expressed as relative percentages % (i.e. 10,6 33,4 28,8 27,2 and 12,4 32,4 27,6 27,6 respectively), verify the equivalence of each samples on the ordinal stage of measurement.

The evaluation on the interval stage of measurement is made with the delivery date of the artist for the time-frequency distribution in 20 intervals of 25 years from 1500 to 1999 (Fig. 1).

Fig. 1

The massive discrepancy between each samples is exceptional. It may be higher proven within the graph of relative frequencies in %, given the totally different sizes of the samples (Fig. 2). Logically, the mixed Pattern (1+2) could be thought of because the much less biased and reveals clearly that the temporal distribution of artists is just not accurately represented in Pattern 1 from interval 1824-50 onwards. 

Fig. 2

Additionally the sharp drop of variety of artists in interval 1900-24 is noteworthy and was not detected in Pattern 1.

Word that this time-frequency distribution relies on the delivery date of the artist and, clearly, doesn’t symbolize the distribution of the full variety of artists (depicting Venus) energetic in every interval. Assuming a coaching interval of 20 years and an energetic life-span of 40 years, the full variety of artists energetic in every interval could be estimated as about double the given figures (5). 

The position of feminine artists

An evaluation on the nominal stage, i.e. the gender, reveals a really small variety of 19 feminine artists in Pattern 1 and a a lot increased variety of 87 feminine artists in Pattern 2

Fig. 3

This time-frequency distribution discloses the position of the feminine artists within the bigger relative variety of artists of Pattern 2 within the 19thand 20th centuries (Fig. 2).


There’s plenty of speak concerning the rising digitalisation of art-historical archives (6). This put up demonstrates with quantitative knowledge the variations between two samples of artists in a selected case associated to the intervals of compilation: 2004-2008 and 2009-2020, respectively. The grouping of each samples leads to a pattern with a smaller bias.

Amongst many accessible WEB developments, one among particular curiosity for this case is the very giant database of the ‘Salons et expositions de teams 1673-1914’with greater than 210,000 notices about artists and their exhibited works (7). Most feminine artists in Pattern 2 have been retrieved from these Salons. 

The total artwork historical past is greater than what’s present in classical artwork historic books or in Google ‘Artwork & Tradition’, particularly with respect to the gender difficulty of artists. It’s time for social decanonization in artwork historical past: “a means of customers’ interactions with cultural content material on the net or by means of digital means (…) with the potential to redirect consideration towards a wider spectrum of cultural manufacturing than was thought of by historic canons.“ (Nuria Rodríguez-Ortega, Might 11, 2018).(8).


  1.  Andrew Piper “There Will Be Numbers” Journal of Cultural Analytics, Might 23, 2016 DOI: 10.22148/16.006
    AndrewPiper, Division of Languages, Literatures, and Cultures at McGill College, is editor of the Journal of Cultural Analytics

    A Digital Thematic Analysis Assortment contributing to distant viewing, macroanalysis and the quantitative strategy in artwork historical past: ‘Venus Iconography’. Six Topical Catalogues have been revealed, with 19,300 artworks by 6628 artists, categorized in 18 foremost subjects.


3. Topical Catalogue ‘The French Venus’ Vol. 2.1  counts 2,997 artworks of 997 recognized French artists and is obtainable at academia.edu
4. The revision of Vol. 2.1 is a long-term undertaking and revealed in subsequent variations. This version 2.1.3 contains now 5,033 artworks of 1,658 artists. The compilation is up to date for 1055 artists with surname within the alphabetical checklist [A,B,C,D]. They symbolize already 64% of the present checklist of artists. 
5. This intrinsic downside has been mentioned in put up of June 29, 2013  Statistics in Art History (V): Drowning in numbers of artists’ 
6. See put up of Might 9, 2020 ‘On-line assets of French artwork for assortment freaks, a sequel’ 
7. See put up of April 27, 2012 Exhibition Catalogues in the Digital Age 
8. See put up of  Might 30, 2018  ‘Meta-Analysisin Digital Artwork Historical past (IV): canonization, decanonization, community evaluation’


    You might also like my different 97 posts since 2011: 

    use ‘search’ or attempt the dynamic view.

    Click on ‘Dwelling’ and pull down the menu of the left button within the header and it’s possible you’ll select one among a number of choices of studying types:

    * Traditional * Flipcard * Journal * Mosaic
    * Sidebar * Snapshot * Timeslide