Modelling the flow of paintings in Venice

Objectives and Deliverables

Maybe you have already been to an art museum at least once in your life and enjoy the some of the artworks there. However, have you ever stood in front of a painting and wondered where the painting had been in the past before it arrived here? Or have you ever wondered what the art collections in the art museum looked like maybe, say, a hundred years ago? And if you saw a painting in the museum a hundred years ago which is not present anymore, would you like to know where it had been since? Our project will try to answer those questions by gathering data not only in the present but also in the past to map the movement of artworks passing by Venice. After that, a web-based instrument might be created to allow user to not only follow the circulation of artworks around Venice, but also allow to visit a specific art museum in Venice in the past. These information will be useful not only for general viewers but also for researchers who would like export useful statistical data of those artworks.


1.Data Gathering

The first and most important step in the project is to gather all the data needed. Those included:

  • The names and locations of the art museums/galleries in Venice (We will prioritize the museums that are still around and had been around long enough, although how long exactly is still to be determined)
  • The names, locations and the digitalized version of the art works that had been passing by Venice. Those included the ones that are in Venice at the moment and also the ones that had been in an art museum in Venice in the past.
  • – The time those paintings arrived in Venice (if they were not there at the first place) and the time those paintings left Venice (if they are no longer there).
  • -The place(s) those paintings had been staying before arriving at Venice and after leaving Venice.

A good way to find the required data is to gain access to the documents in each art museum in Venice where the records of the paintings might be recorded. Further in the project, we will see clearer how much data is available to be collected.

From the information above, it is obvious that the size of the data required is huge. This is why it is necessary first to narrow down the scope of the project to fit the time given. For the time being, we temporarily limit the number of history locations data of a painting to up to 3 points with the museum in Venice in the center. This is demonstrated in the sketch below:


Hence, for every paintings in or used to be in Venice, we will obtain up to 3 locations data and 2 temporal data being the time it arrived and left Venice. The data, however, still required to be downsized, for example, we can focus only on a specific time frame in the past (the Renaissance, for instance) or to a specific numbers of museum in Venice. However, those decision can only be made once we understand the limit in the data that we can collect.

Another remark is that, for the time the painting stayed in Venice, we are going to give the exact name and position of the museum that stored it, however, this is not always the case for the other two locations unless both of them are also in Venice. This is because if these two remaining locations are from other city or even overseas, mapping exactly the name of the museums would similarly mean that we have to map the museums outside of Venice as well and this would be impossible for the time given. Hence, we’ve decided to limit the precision to the name of the city only. For example, painting “A” stayed in city “C1”, then in 19XX moved to museum “M” in Venice, then in 19YY moved to city “C2”.

2.Data Mapping

A map of Venice with names and locations of the museums of interest (to be decided) should be produced from the existing data (called map1). Then this map of Venice will be matched with a larger typical map of the world (map2), with the names and locations of all the countries and cities. Map2 hence, can be borrowed directly from an existing digital map online. After that, the data will be imported to points on the map. The data included the location data, time data, and also the digitalized version of the paintings.

By doing this, museums, galleries and other concentration of work of arts could therefore be explored as they once were by accessing a single location point on the map. (for example, we will have at location (x1,y1) the list of paintings which have stayed at that point)

3.Data Categorization

Any work of art, and particularly any painting, is subject to many types classification. Because of this, it would be very interesting if we can organize all the paintings within the database by a specific classification, say, by artists, genres, styles… To do this, a painting should be connected to many tags, each tag refer to a type of classification. For example, The Mona Lisa would be tagged by “Leonardo Da Vinci”, “Portrait”, ”Renaissance”…

One would then be able to search on these paintings using the tags defined for each painting. To this end we need to implement 3 operations :

  • Union (+) : Monet + XIX would yield all the paintings by Monet and all the paintings from the 19th century.
  • Intersection (*) : Monet * XIX would list out everything Monet painted in the 1800s that passed through Venice.
  • Difference (-) : Monet – XIX would, for instance, list what Monet painted in the 1900s.

This list of operations guarantees that any painting, and any set of tags can be easily accessed. If possible, we can also allow users to add tags into paintings but this would in turn reduce the precision of the information.

Finally, some tags can be exported from a painting using simple signal processing command. If we can realize this, one could type in the search bar “Bright” and the program could give them all the result of all the paintings with the bright tone. This could allow users to find paintings directly based on their image data.


The final product will be made browser based in order to make it available to all users online. Using which the users can make use of the cursor on the map and/or the search bar to access to paintings the way they want. This would allow them to find paintings based on their preference or simply allow an user to visit a museum in Venice in the past. In order to have a large amount of data we can include a feature so that users can add tags and information to the map, however these should be marked as “unverified” information.


  • Gathering data & narrow down the scope of the project        2-3 weeks
  • Import the data                                                                                         2 weeks
  • Mapping (museums and flow of paintings)                                   4 weeks
  • Categorization (with tags and search language)                         3 weeks
  • Polishing, creating web interface (if possible)                              2 weeks
  • Image processing features                                                                    If time permits