The article presents possibilities and limitations of quantitative analysis of databases of visual data (in thiscase it is analysis of database of photos). The axis of the article is analysis of database of “Invisible city”project (web page: http://www.niewidzialnemiasto.pl/), its first goal was photographic documentation ofgrassroots forms of modification of city space done by inhabitants of large Polish cities (broadly speakingthis informal activity reduces to beautification of life environment and to improvement of its functionality).The moment the database of “Invisible city” project had grown to the size of several thousands of pictures, the idea arose that collected material can be analyzed in a quantitative way, which would enable more general, systematic view on phenomenon interesting for researchers.In the article, firstly, there are presented tools, scheme and general results of analysis of photos gathered indatabase of “Invisible city” project, here, in other words, is presented the scheme of transformation ofvisual data to numerical data. Secondly, there are identified moments where the scheme of quantitativeanalysis of “Invisible city” fails, where it may result in information loss, doesn’t give us one hundredpercent certainty for obtained results (problems of first kind are connected with dynamics of the formationand growth of database, problems of second kind arise at the stage of pictures analysis). Thirdly, there isproposed and described one potential master (meta)analysis, that we can impose on analysis of databaseof visual data of “Invisible city” project in order to check its accuracy. Despite of mentioned above problems,difficulties associated with the fact that database of “Invisible city” was created spontaneously to some extent (it was balancing between database and photo gallery sometimes), results of quantitative analyzes are reliable, they reconstruct the dynamics of invisible city with accuracy and give strong foundation for in-depth description.
Read full abstract