Art Mapping Venice Abstract

1. Introduction

The coming of digital era brings us new tools to come up with solutions for humanities problems. These tools in general make the solution a lot easier. This is the case for the geographical information systems. Geographical information systems are the tools that enables to capture and make use of geographical data for different purposes. These tools make dealing with geographical data a lot easier. For example, Jacopo de’ Barbari, an Italian painter, mapped the city of Venice in great detail in late 14th century. This should have taken a lot of effort at that time to map a city in great detail. However, if he lived in this age, the same task for him would be a lot more easier to realize with the help of digital tools.

Geographical information systems opens up new areas of research one of which is art mapping.  Art mapping is to represent an artwork by making use of associated geographical information of the artwork on a map. This way the user can easily associate the spatial information with the artwork. For example, Art Mapping Venice project realized by Orhan Öçal and Tania Palmieri represents the city of Venice by its art works. Since my project is to extend Art Mapping Venice project, it would be a good idea to briefly explain this project.

The purpose of the Art Mapping Venice project is to map the city of Venice through art works such as paintings, photos, drawings. In order to do this, they created a user friendly web-page where you can see the current city map with aerial shades on top of the historically or artistically important places which they call as point of interest. The user can browse through the map, zoom in and out. This map is just like what we see on Google Maps since they have used Google Maps interface and some of its functionalities while realizing the project. When the user clicks on those areal shades, they are provided with a small pop-up window where the user can see the artwork associated with the place. Moreover, when the user clicks on information button he/she is provided with more information about the place, paintings together with contemporary photographs of that place.

old projectArt Mapping Venice Web Tool

2. Research Question

The research question I would like to solve is to extend this web tool by adding the functionality that user can switch between an historical map of Venice and current map of Venice. In the current version of the project, old representation of a specific point is shown on a small pop-up window. It would be a good extension to show the old representation as a whole map or some part of the map with the same orientation . This way user can see the differences between the two maps and see the changes over the course of history.

In order to realize this project, first question would be to find an old map of Venice that has a good quality and is easy to match with the current map of the Venice. For that matter, Jacopo de’ Barbari’s map of Venice published around 1500 is used. This map not only has a good quality but also covers the area that we are interested in. The next question would be about identifying point of interests on the old map and matching them with the current map. Moreover, when the user switches between the old map and current map, the orientation of the maps should match so that user can compare the old map and current map easily. Another challenge is about representing the old map on a geographical information system. It would be a good idea to continue with Google Maps interface since the existing project uses it. For that reason, how to represent an image on Google Maps interface is also a part of the research.

In brief, the goal is to add user-friendly functionality that user can switch between an historical map of Venice by Jacopo de’ Barbari and current map of Venice to the Art Mapping Venice project. Achieving this goal requires going through the following research phases:

– Identification and representation of point of interest locations on the Jacopo de’ Barbari’s map

– Representing Jacopo de’ Barbari’s map on Google Maps

– Transition between Jacopo de’ Barbari’s map and current map of Venice

3. Identification and representation of point of interest locations on the Jacopo de’ Barbari’s map

In order to realize the project, it is necessary to identify the point of interest locations on the Barbari’s map. Those places are basically historically important places of the city of Venice. For that reason, it is likely that those places are also covered in Barbari’s map. There are 20 point of interest locations on the existing project. After doing some research, it turned out that 14 of those places are also covered in the Barbari’s map.  These places are San Geremia, Ponte di Rialto, Grande Canal,  San Giorgio Maggiore, Piazza San Marco, Scuola di San Marco, Entrance to Grand Canal, Prigioni, Fondaco dei Turchi, Arsenal, Palais Contarini, Ca d’Oro,  Santa Maria Gloriosa dei Frari, Palazzo del Michiel.  The only thing to note here is that Fondaco dei Turchi does not look the same since it has undergone thorough restorations between 1860 and 1880. 6 remaining places were not identified because of the following reasons. Osteria dei Sabbioni, Villa Foscari are outside the area where Barbari’s map covers. Santa Maria della Salute (built in 1687), Teatro La Fenice (built in 1792),  Palazzo Dolfino (built in 1536), Église du Rédempteur (built in 1592) are built after Barbari draw the map.

In the existing project, point of interest locations are represented by polygon drawings. This is done by preparing a KML database to store the information and process it to show the drawings. Similar approach is adopted for the Barbari’s map also. In order to show the polygon drawings on top of point of interest locations, KML database corresponding to point of interests on the Barbari’s map was prepared. The coordinates of the corners for each of the point of interest is found out and written to the KML database. Since for the KML database everything will be the same as the existing database except for the coordinates, the same KML database is used and only the coordinate values are changed.

piazzasanmarcoPiazza San Marco on Jacopo de’ Barbari’s Map with Polygon Drawing on Top

4. Representing Jacopo de’ Barbari’s map on Google Maps

In order to show the Barbari’s map on Google Maps, two possibilities were found out. One possibility is to use define a different map type for the corresponding image by using image map types documentation of Google Maps API. Second possibility is to use adding a custom overlay example on Google Maps API documentation.

In order to explain the first approach, it is necessary to explain the tile concept on Google Maps. While Google Maps is loading, instead of loading the whole map, it actually loads square small parts of the map in order to make the processing faster. These small parts are called tiles. For defining the image map type, the idea is to provide an image source for each tile. For that matter, it is necessary to break the Barbari’s map image into pieces beforehand so that we can provide image source file for each tile. Also, note here that these images change for different zoom levels. Therefore, we need to break the Barbari’s image into pieces for each zoom level and store them. This approach is inefficient in terms of storage. Moreover, this approach brings difficulties in terms of implementation.

tilesGoogle Maps Tiles

In adding a custom overlay example, you specify the coordinates of the corner points where the image is put. Then, whatever the zoom level is, the image stays at the same location. Since this approach does not require us to worry about the tiles, the implementation becomes simpler and there is no need for storing chunks of images for each zoom level and each tile. The only problem with this approach is that the loading of the image is slow since Google Maps tries to load the whole image.

Considering its advantages, adding a custom overlay example is used for this project.

5. Transition between Jacopo de’ Barbari’s map and current map

An important phase of the project is to decide on how to make a transition between the Barbari’s map and the original map. As mentioned earlier, 14 point of interest locations were identified on Barbari’s map. This means that we can easily find center coordinates for the 14 of the point of interest locations on both Barbari’s map and original map. Based on those center coordinates (call reference points) and given a chosen point (call linking point), the following transition method between the two maps is proposed.

The idea is to find out the coordinate of the corresponding linking point on the map we want to switch to. To achieve this task, we first calculate the distance between the linking point and reference points on the current map. Then, each corresponding reference points on the other map is translated with the calculated distance in the same direction. Then, we obtain a set of points. Here one idea would be to find the center of those points. However, this wouldn’t give a good result since for the reference points that are far away from our linking point should not have much effect compared to the ones that are close. In order to reduce the effect of the points that are far away, center of mass is calculated by assigning small weights for the reference points that are far away and large weights for the reference points that are close. This weighting strategy is determined experimentally. Once we have the corresponding linking point on the map we want to switch to, we can show the corresponding map by just overlapping the two linking points.

6. Results

The project was successfully completed with the following features:

– Jacopo de’ Barbari’s map is added as a top layer image on top of the original Google’s map.

– The user can see the point of interest locations (14 of them) on the Barbari’s map with aerial shades on top of those places. When the user clicks on those aerial shades, the same pop-up window with same information in the existing project is shown.

– The user can switch between the original map and the Barbari’s map given a reference point by simply clicking the “Change Map” button. The location of the reference point can be changed so that it is possible to see how the transition is affected by the location of the reference point.

– The user can change the transparency of the Barbari’s map by using the slider on the web interface.

Screenshot from 2013-12-23 15:15:27

Art Mapping Venice Web Tool with Jacopo de’ Barbari’s Map Added

All in all, the desired functionality were added to the existing project without changing the overall design much. With the help of added functionality, it is now possible to compare the current map of Venice and map of Venice 500 years ago. The user can see how the city of Venice changed over 500 years. Moreover, the user can also see the effect of having a panoramic map rather than having top view map. Barbari’s map is a panoramic view map. That is why the places that are at the south are drawn bigger compared to the places that are at the north. This difference actually makes it difficult to match the two maps.

7. References

[1] Jacopo de’ Barbari. Web. http://en.wikipedia.org/wiki/Jacopo_de%27_Barbari Date of access: 23 Dec 2013

[2] The Journey of ArtMapping Venice. Web. https://artmappingvenice.wordpress.com/2013/05/11/the-journey-of-artmapping-venice/ Date of access: 23 Dec 2013

[3] Fondaco dei Turchi. Web. http://en.wikipedia.org/wiki/Fondaco_dei_Turchi Date of access: 23 Dec 2013

[4] Villa Foscari. Web. http://en.wikipedia.org/wiki/Villa_Foscari Date of access: 23 Dec 2013

[5] Santa Maria della Salute. Web. http://en.wikipedia.org/wiki/Santa_Maria_della_Salute Date of access: 23 Dec 2013

[6] La Fenice Theatre. Web. http://en.wikipedia.org/wiki/La_Fenice Date of access: 23 Dec 2013

[7] Église du Rédempteur de Venise. Web. http://fr.wikipedia.org/wiki/%C3%89glise_du_R%C3%A9dempteur_de_Venise Date of access: 23 Dec 2013

[8] Google Maps JavaScript API v3 Documentation. Web. https://developers.google.com/maps/documentation/javascript/basics?hl=tr Date of access: 23 Dec 2013

[9] Google Maps Adding a Custom Overlay. Web. https://developers.google.com/maps/documentation/javascript/examples/overlay-simple Date of access: 23 Dec 2013

[10] Google Maps Image Map Types. Web. https://developers.google.com/maps/documentation/javascript/maptypes#ImageMapTypes Date of access: 23 Dec 2013

Advertisements

Art Mapping Venice: Final Steps

This blog post is to show you the progress made for the Art Mapping Venice project during the time after last blog post. As mentioned in the last progress blog post, the Barbari’s map is added to Google Maps interface as another layer on top. In order to continue with the rest of the project, the location of the Barbari’s map is fixed. This is necessary since we need the determine the coordinates of the point of interests on the Barbari’s map so that we can draw polygons on those places. However, this does not mean that the location of the map will stay the same. The location of the Barbari’s map will be determined by the proposed method which is also explained in this blog below. The important point here is to match zoom levels of the two maps. However, since Barbari’s map is not a top view map, it is not possible to match the zoom levels correctly. For that matter, the following approach is adopted. The aspect ratio of the image is kept the same. Then, the Grand Canal on two maps are more or less overlapped.

Next step is to identify the point of interest locations on the Barbari’s map. On the existing project, there are twenty point of interest locations. However, only 14 of those places were identified. These places are San Geremia, Ponte di Rialto, Grande Canal,  San Giorgio Maggiore, Piazza San Marco, Scuola di San Marco, Entrance to Grand Canal, Prigioni,  Arsenal, Fondaco dei Turchi, Ca d’Oro, Palais Contarini, Santa Maria Gloriosa dei Frari, Palazzo del Michiel.  The only thing to note here is that Fondaco dei Turchi does not look the same since it was rebuilt around 1800s. 6 remaining places were not identified because of the following reasons. Osteria dei Sabbioni, Villa Foscari are outside the area where Barbari’s map covers.  Santa Maria della Salute (built in 1687), Teatro La Fenice (built in 1792),  Palazzo Dolfino (built in 1536), Église du Rédempteur (built in 1592) are built after Barbari draw the map.

In order to show the polygon drawings on top of point of interest, KML database corresponding to point of interests on the Barbari’s map was prepared. Since the point of interests are identified, the coordinates of the corners for each of the point of interest is found out and written to the KML database. Since for the KML database everything will be the same as the existing database except for the coordinates, the same KML database is used and only the coordinate values are changed.

An important phase of the project is to decide on how to make a transition between the Barbari’s map and the original map. Here the proposed method for transition is explained. Center coordinates for the point of interest locations on both Barbari’s map and original map is known. Based on those center coordinates (call reference points) and given a chosen point (call linking point), the following transition method between the two maps is proposed. The idea is to find out the coordinate of the corresponding linking point on the map we want to switch to. To achieve this task, we first calculate the distance between the linking point and reference points on the current map. Then, each corresponding reference points on the other map is translated with the calculated distance in the same direction. Then, we obtain a set of points. Here one idea would be to find the center of those points. However, this wouldn’t give a good result since for the reference points that are far away from our linking point should not have much effect compared to the ones that are close. In order to reduce the effect of the points that are far away, center of mass is calculated by assigning small weights for the reference points that are far away and large weights for the reference points that are close. This weighting strategy is determined experimentally.

Having explained the details of the progress made, it is now time to show the most recent view of the web interface.

shsproje

Screenshot of the web interface of  Art Mapping Venice

In the above picture, the polygon drawings on the point of interest locations are shown. Moreover, there is a red marker object on the map. This marker is our linking point which is used when we want to switch from one map to the other. The user can change the location of this marker object to see how transition is affected for different linking points.

 The plan of the following weeks is to:

– Finalize the project

– Write the final report

– Prepare the presentation

Art Mapping Venice: Progress and Implementation Choices

While making progress towards the realization of the Art Mapping Venice 1.1 project, I got familiar with the existing project. I used the tool extensively since this project requires the use of existing project and add some functionalities that was mentioned in the previous blog post. This extensive use enabled me to think about how to add the functionalities to the existing project without changing overall design too much. This blog post not only explains what have been done so far but also gives a roadmap of what will be done in the next steps towards the realization of the project.

The very first challenge was to find a historical map of Venice that has a good quality. For that we were able to find the Jacopo de’ Barbari’s map of Venice in good quality image. The production of this map dates back to 15th century. At the end of the project, the user will be able to compare the current map of Venice with its map 500 years ago, which sounds really cool.

Jacopo_de'_Barbari_-_Venetie_MD_-_retouched

Jacopo de’ Barbari’s Map of Venice [1]

Another challenge faced during the first phase of the project was about georeferencing issue. Let me first explain what georeferencing is and then talk about its effect on Art Mapping Venice project. According to ESRI, georeferencing is the process of “aligning geographic data to a known coordinate system so it can be viewed, queried, and analyzed with other geographic data” [2].

In our context, we were planning to align the image to the Google Maps coordinate system. This is done by matching a set of points on the original map of Google and the map that we were planning to georeference. We used ARC GIS software for that. However, we were not able to get a good result after some trials. This is because the two maps differ a lot. For example, even if we use three reference points for georeferencing, the picture was deformed a lot. After trying different number and set of reference points, we come to the conclusion that it will not be useful for us to georeference the image. We decided to move on with the project without worrying about georeferencing. This actually brings us another implementation choice we made that will be explained later in this blog post. Although georeferencing will not be useful for the final output of the project, it was good to learn the concept of georeferencing and also to see how much the old map differ from the current map.

After coming to the conclusion that we won’t be able to georeference the Barbari’s map, we decided to use the Barbari’s map as it is as a layer on top of the usual map of google. However, since these two maps differ a lot, we needed to think of a solution to have a good representation of the old map. The solution proposed is to determine the coordinates where Barbari’s map is put will be determined by a mapping given a reference point. We will match the reference point on the old map and the current map. Since it is not possible to do this matching for each and every point on the map, we decided to do this by looking the location of the reference point with respect to point of interests. Since we will identify these point of interest on the old map and they are already identified on the current map, we can use them for the transformation. How those points are going to be used will be the topic of further research in the forthcoming weeks.

The next phase of the project was mostly about doing research about technical details. Since at the beginning, I didn’t have a good insight about how to overcome technical challenges, I proposed first to focus on collecting data and metadata that will be necessary. However, then I realized that it would be a better idea to focus first on implementation details. This is because I needed to make sure that it is possible to use Google Maps platform to realize the project. If it wasn’t the case I needed to find some other platform and that would be a big problem. That is why I decided to first focus on implementation details first. After doing some research, it turned out that it is possible to continue using Google Maps platform since Google Maps allows us to use an image on top of its own map. This will enable us to use Barbari’s map as another layer on top of the usual Google’s Map.

Screenshot from 2013-11-25 11:49:36

Jacopo de’ Barbari’s Map Added as Another Layer

Following is the plan of what will be done in the next steps:

– Preparing KML layer entries necessary for the old map so that drawings on the old map can be shown.

– Getting the center coordinates of the point of interests on the old map and the current map. This coordinates will be used for the mapping of the coordinates while switching the maps.

– The mapping of the coordinates when switching from one map to the other given a reference point.

Art Mapping Venice v1.1

1. Introduction

The Art Mapping Venice project attracted considerable attention among both digital humanities scholars and EPFL community. This project won silver award in the DH101 course last year. Moreover, there was a post on the main page of the EPFL website and EPFL community talked about this project. These all show that this project has made a big success. First of all, I would like to thank you my friends Orhan and Tania for their success in realizing Art Mapping Venice project and letting me working on extension of the project.

EPFL

Art Mapping Venice on EPFL Mediacom

The research question that I would like to solve is to extend this Art Mapping Venice project by adding some more functionalities. That is why it would be good idea to explain briefly what has been done in this project last year and what was the final output. After all, understanding the details of the Art Mapping Venice project will be an important part of the work while extending the project this semester.

The purpose of the Art Mapping Venice project is to map the city of Venice through art works such as paintings, photos, drawings. In order to do this, they created a user friendly web-page where you can see the current city map with areal shades on top of the historically or artistically important place which they call as point of interest. The user can browse through the map, zoom in and out. This map is just like we have on Google Maps since they have used Google Maps interface and some of its functionalities while realizing the project. When the user clicks on those areal shades, they are provided with a small pop-up window where the user can see the art work associated with the place. Moreover, when the user clicks on information button he/she is provided with more information about the place, paintings together with contemporary photographs of that place.

artmapping

Art Mapping Venice

2. Research Question

The research question I would like to solve is to extend this web tool by adding the functionality that user can switch between an historical map of Venice and current map of Venice. In the current version of the project, old representation of a specific point is shown on a small pop-up window. It would be a good extension to show the old representation as a whole map or some part of the map with the same orientation without opening a new page. This way user can see the differences between the two maps and see the changes over the course of history.

In order to realize the project, first question would be to find an old map of Venice that has a good quality and is easy to match with the current map of the Venice. One possibility would be to use Jacopo de’ Barbari’s map of Venice published on 1500. However, it is necessary to do some research to find the most appropriate map for the project. The next question would be about identifying point of interests on the old map and matching them with the current map. Moreover, when the user switches between the old map and current map, the orientation and zoom of the maps should match so that user can compare the old map and current map easily. One idea would be to match the zoom and position of the chosen point of interest in the old and current map. Another possible challenge is about representing the old map using Google Maps interface. We do not know if Google Maps has functionality to put different map instead of its own map. What could be done if Google Maps does not allow such a functionality will require some research to be done to find other ways of representing the historical map.

In brief, the purpose is to add user-friendly functionality that user can switch between an historical map of Venice and current map of Venice to the Art Mapping Venice project.

3. State of the Art

Geocoded Art, MOCA’s Ends of the Earth, TATE’s Art Maps, History Pin are the examples of the possible realizations of art mapping of a city. All these projects were discussed earlier that can be found on Art Mapping: visualizing paintings in real world using digital maps blog post.

Although these projects give an overview of possible art mapping representations of a city, the extension of the project is somewhat different in the sense that it will enable user to go from the current map of a city to corresponding historical map of the city or vice versa. The following two projects are more related to the research problem I would like to solve.

The first project is called Manchester Historical Maps, which can be reached on the link:  Manchester Historical Maps. This is a possible representation of a city with its map by providing the user with the current and historical map of a city. On this web-page, the user is first provided with current map of the city of Manchester. On the right side, the user can change the display of the current map with options such as satellite view, open street map etc. The user also has the option to choose an historical map from the list. When the user clicks on a old map that he/she likes to see, the old map is put on top of the current map of the city.  Then the display is like two layers on top of each other. The user can browse through the map and zoom in and out. The user can also choose different transperancy options for the map that is put at the top.

Screenshot-Manchester Historical Maps - Mozilla Firefox

Manchester Historical Maps

Second project is called A Vision of Britain Through Time, which can be reached on the link:  Old Maps of Britain and Europe. This project is another visualization tool that enables user to see the map of Great Britain on the 19th century, 20th century and currently. The user can choose from the list which map he/she wants to see. Moreover, user can zoom in and out browse through the map. This tool is more professional in the sense that they have good quality map of a very large area. However, it lacks the satellite view functionality.

Screenshot-Old maps of Britain and Europe from A Vision of Britain Through Time - Mozilla Firefox

Old Map of Britain by A Vision of Britain Through Time

4. Methodology

There are five milestones of this project. First, it is necessary to understand and get familiar with what has been done until now. This part will mostly focus on understanding the code and read the documentation. Second milestone will be data preparation. This part includes finding historical maps of the city of Venice and preparing the metadata associated with it. Next step would be to build the database that is also compatible with the existing project. Final step would be to design the user interface. The user interface is not going to change completely but it will be necessary to make some modifications which will require some work.

Getting familiar with what has been done

In order to start the project, it is required to have an understanding of what has been done. To get this understanding, it would be good idea to read all the documentation provided. Moreover, in order to understand implementation details of the project, it would be necessary to do some research so that I can have sufficient background to understand the code. This research will include starting learning Google Maps API, XML (which is used for artwork entries), KML(which is used for geographical entries). This part also includes getting familiar with the existing data and metadata associated with it.

Data Selection

This step will require some research to find a suitable historical map for our purpose. To be more precise, here is the explanation of suitable map:

– Has good quality

– Covers the whole area where we have point of interests

– Easy to identify point of interests on the map

We have the Jacopo de’ Barbari’s map of Venice as a candidate. However, it is necessary to check if that map meets our requirements. If that is the case, we can continue with the rest of the milestone since we have the good quality images from Jacopo de’ Barbari’s map of Venice. This part also includes preparing the metadata.

Building Database

At this point, we do not know if we will have a single image file or a collection of files. If we have just a single image file, we might not need a database. However, if we have a collection of images, then it would be necessary to build a database. To do that, XML (Extensible Markup Language) is a possible candidate since they have chosen to use XML for storing some of the data in the project. However, other options like JSON are also going to be considered.

User Interface Design

For the user interface design, the important point is if Google Maps allows us to put an image on top of the its map or instead of its map. The first thing to do is to figure out if Google Maps allow that. If Google Maps does not allow that we need to think about some kind of tricks. The rest of the user interface design will be about making some changes on the overall design so that the functionality to switch between maps will also be included. However, in order to do that i need to have more insight on the libraries and the code they have used for user interface design.

5. Project Plan

1- Writing the Proposal (27.09.2013)

2- Getting familiar with Art Mapping Venice 1.0 (11.10.2013)

2- Doing research on best possible map of Venice (18.10.2013)

3- Identification of the point of interests on the old map (25.10.2013)

4- Preparation of data and meta-data(01.11.2013)

5- Getting familiar with Google Maps API (08.11.2013)

7- Doing reserch on how to integrate functionalities of Google Maps with historical map (15.11.2013)

9- User Interface Design (29.11.2013)

10- Beta Implementation (06.12.2013)

11- Webpage Implementation (13.12.2013)

12- Documentation of the project and Presentation (20.12.2013)

The Journey of ArtMapping Venice

1. Introduction

The coming of the Digital Era has brought irreversible changes in everyday life, introducing habits and methods that one couldn’t have imagined thirty years ago. One of the major changes is without any doubts the availability of the Internet almost everywhere, which brings the ability to access almost any kind of information anytime.

One of the results of this digital revolution related to our project is the printed maps on paper giving their way to extensible online digital maps and navigators. For this reason, in the last years we have seen an increasing development in what are called Geographic Information Systems (GISs), tools designed to capture, store, manipulate, analyze, manage, and present various types of geographical data. The potentials of GIS’s are enormous: the concept of map evolves by the integration of different types of information giving multi-dimensionality that pushes the limits of what can be done on a printed paper map. If the supplementary dimension added to the map is the time and the main represented data are the artworks, the process starts to become interesting by itself.

2. ArtMapping

Our first research question constituting the basis for the development of the project is represented by the definition of the term ArtMapping. Our idea stems from the desire to revolutionize the conventional approach for the discovery of a city, enriching this experience by exploiting the possibilities offered by digital maps. The means of doing this is to merge the details offered by an art catalogue with the spatial information included in a map, in order to associate the artworks depicting a particular subject in the city with the corresponding location today. Starting from these consideration, what we define with the word ArtMap is a tool which provides both the information of a map and those contained in an art book and visualize them together in a user friendly interface.

The city that has drawn our attention more than the others for ArtMapping is Venice. Playing a central role in Renaissance and being one of the essential destinations of Grand Tour, a 17th century tourism boom within aristocratic Englishmen, Venice has extensively been the subject of artists. Due to these two factors, it is relatively frequent that one finds the veduta, highly detailed painting of a city scene, of sceneries of Venice by different artists done in various periods. In this context it is especially important to remark the painters Canaletto (1697-1768) and Michele Marieschi (1710-1743), who are among the most famous vedutisti, veduta painters, of their periods.

In order to design and realize the ArtMap of Venice, we needed to define and choose the necessary ingredients. These elements fundamentally consist of:

  • The works of art: which subcategory among the several possibilities, such as prints, paintings, photos, drawings, etc., will constitute the collection that is integrated in the ArtMap.
  • The GIS: which platform is used to show the geographical information associated with each location and its related artworks.

3. Collection of artworks: Merveilleuse Venise

As the source of the artworks of Venice, we decided to focus on a particular book, “Merveilleuse Venise” by Sophie Monneret, published in 1971. The book contains prints of Venice dating from 15th century to 18th century, drawn by famous vedutisti including Canaletto and Michele Marieschi, that are preserved in the collections of Bibliothèque Nationale de Paris and Musées Civiques d’Art et d’Histoire de Venise. The drawings that are found in the book are of various types including city plans, vedute, depiction of public events and daily life in Venice. This variety of nature of drawings is especially suitable for our project, since it allows us to capture the history of Venice more thoroughly. Furthermore, the drawings are accompanied by explanatory text relating the drawing to history, which is again a big plus for conveying the story of Venice.

Ponte di Rialto, being built in 1591 it still stands as one of the most frequently used bridge of Venice.

Ponte di Rialto, built in 1591, still stands as one of the most frequently used bridge of Venice.

Mapping the prints from one single book is beneficial for our work, allowing us to collect data in a systematic way and making use of a reliable source of information, which is very important for achieving a reliable temporal resolution of Venice in the ArtMap. 

3. The geographical interface

The next decision we needed to make in order to realize our project, after defining the source of the artworks, was choosing the appropriate geographical interface that is able to relate the point of interests in Venice with the historical prints. There were a number of options to choose from for this interface, three major ones being Google Maps [4], Google Earth and ArcGIS where each one had their pros and cons.

After testing out the capabilities and usability of different interfaces, we have chosen the Google Maps interface over the others since it meets the needs of our project while being easy to use in the end users side, not needing any plug-ins to be installed. Furthermore, with this interface the user has the ability to view the city as a map or as registered satellite images. While map view gives an easy to grasp and clean representation of the city, thus accentuating the information in the prints, the satellite view gives detailed scenery information enabling the user to immerse themselves more in Venice. Furthermore, by having automatic support for aerial close-up images and changing the point of view, Google Maps interface enables the user to discover the point of interest from four different angles, giving them the ability to match the desired angle of view, in particular with the artists’.

Piazza San Marco looking north.

Piazza San Marco looking north.

Piazza San Marco looking east.

Piazza San Marco looking east, which meets the point of view of the artist more.

4. The database for point of interests and images

Since the aim of our project is to create a spatio-temporal representation of Venice, we present the user two fundamental information: geographical information of locations of Venice and their evolution with time visualized through historical prints. In order to meet these two goals, the information of locations and the prints need to be stored in a database. Two main ways this can be done are by storing the information of locations and the prints in the same database or splitting the two and relating the information between these two databases. For reasons of standardization of the geographical information used in the project and utilizing powerful representation abilities of Keyhole Markup Language (KML) without violating its standards, we decided to choose the second way and built two separate databases, one for representing geographical locations and one for the prints.

Databases

The geographical (KML) and artwork (XML) entries for Scuola di San Marco. The relation between two databases is made through the matching of id tag in KML’s description and Point of Interest (<poi>) tag in XML.

There are a number of benefits of using standard KML. First of all, KML is very powerful in terms of geocoding, having capabilities of representing complex shapes, view angles and geographical information. One of the fundamentally important benefit KML brings to our project is that it can represent polygons of different colors and styles, which enables us to confine the point of interests, buildings, piazzas, canals, etc., precisely and classify them visually according to their type by styling the polygons. Furthermore, by implementing the KML support using geoxml3 API [5] which is designed for customized interaction with placemarks in the database, the user can hide the polygons for removing the transparent overlays for better visualization of the point of interests. Next, KML is a proliferated and well supported standard for representing geographical data, which makes the completed database independent of the visualization tool. The same KML data can be imported to GIS software and the data complying with the KML standards is displayed without any special programming being necessary. Furthermore, using a known standard for representing information creates the possibility for other enthusiasts to make databases that can be displayed in our interface.

For storing the information on the prints of Venice we chose to use an XML file. The XML file stores the name of the print, the artist, the date of the print, description of the location, the tag of the location and the image files of the print and a contemporary photograph that approximating the point of view of the artist. The historical information about the location is queried when the user selects a location from the Google Maps interface. The gathered information from the XML file about the prints matching their location tags with the ID of the selected point of interest is then showed in a modal window [6]. With this implementation, multiple images corresponding to single location can be showed, enabling the user to see how the location evolved through time.

Fondaco dei Turchi has been undergone thorough restorations between 1860-1880 during which two towers were added.

Fondaco dei Turchi has undergone thorough restorations between 1860-1880 during which two towers were added.

5. ArtMaps features

The artmap final interface is designed in order to offer the user the primary tools needed for the artistic discovery of Venice:

  • The underlying map, with the possibility to switch from map mode to satellite or  earth mode, adapts to user preference.
  • The point of interests, that are well visible with every visualization mode on the map and on an external list, are coded with different color codes for churches, buildings, squares and cultural events allowing the user to quickly focus on the specific target.
  • The modal window, in which the historical prints are visualized side by side to the current appearance of the location [7], provides an easy interface for following the history of Venice.
Entrance to Venetian Arsenal: Porta Magna

Entrance to Venetian Arsenal: Porta Magna

Particular importance is given on the concept of temporal resolution: when there are more than one print available for a single location, the modal window displays the prints in chronological order, with the corresponding present appearance matching the point of view of the artists’. In this way, the user is given the possibility to see and compare the different drawings, realizing the changes that the particular point of interest has undergone during the years.

Besides the visual information, a brief description is available for every location in order to summarize the related history and curiosities. Furthermore, external information such as links to related webpages or embedded videos are included for extending the presentation of the location.

An embedded YouTube video about Festa del Redentore.

An embedded YouTube video about Festa del Redentore.

6. Conclusion

We have developed a means for visualizing how Venice changed through time by using artworks. The interface is designed to be user friendly, where the historical information is visualized on a digital map. As the user requests information about a location, they are supplied with prints of Venice dating from 15th to 19th century and contemporary photographs of the same location for comparison. Furthermore, besides the visual information about the location, the user is given a description of the history and the importance of that particular place in the history of Venice. With this property, the tool is also suitable for touristic visitors of Venice for gathering visual and historical knowledge about the city and its important landmarks.

Although the designed website tells the story of Venice captured by the book “Merveilleuse Venise”, it is easily extensible since the geographical information is represented in widely accepted standard KML and the image database represented in XML follows simple tagging rules. By extending only these two files, the tool can be used for visualizing other cities or other landmarks retreived from different sources.

To conclude our project, we think ArtMapping is a user friendly concept for telling the history and helping the discovery of a city by using the artwork of famous artists of the past and visualizing them through the modern mapping tools of the present.

References:

[1] Venetian School. Web. http://en.wikipedia.org/wiki/Venetian_School_(art) Date of access: 14 May 2013

[2] Venice Art History. Web. http://www.aboutvenice.org/art-history-of-venice.html Date of access: 14 May 2013

[3] GIS. Web. http://en.wikipedia.org/wiki/Geographic_information_system Date of access: 14 May 2013

[4] Google Maps JavaScript API v3. Web. https://developers.google.com/maps/documentation/javascript/ Date of access: 14 May 2013

[5] Geoxml3. Web. https://code.google.com/p/geoxml3/ Date of access: 15 May 2013

[6] SimpleModal. Web. http://www.ericmmartin.com/projects/simplemodal/ Date of access: 14 May 2013

[7] Panoramio. Web. http://www.panoramio.com Date of access: 14 May 2013

Final steps towards the realization of ArtMapping Venice

In the previous blogs we have talked about the source of the historical paintings, the book Merveilleuse Venise  and what platform we are going to use to display the geographical information. After deciding on using Google Maps as the interface of our project, next step was to decide on the format of  the database for storing the geographical information. In our first blog post we mentioned that we set up a simple, specially designed JSON file containing both the geographical information and information about drawings of Venice in the same place. By implementing domain specific parsing and visualization functions to work with this JSON file, we set up a template homepage to show the geographical and painting data on Google Maps for testing its capabilities and seeing which tools of API can be useful for the realization of our project.

Besides using a specially design JSON file, another option for the creating database for geographical information was to use Keyhole Markup Language (KML). At first we were more inclined towards using the JSON file that we would design to accommodate our primary need which is to build a spaciotemporal representation of Venice extracted from the book Merveilleuse Venise. However, as a result of long discussions and observation of the powerful geocoding capabilities of KML by implementing a test database, we have decided to represent the geographical information by using KML.

Homepage using Google Maps and KML

Homepage using Google Maps and KML

Advantages of standardized KML over specially designed JSON file.

There are a number of benefits of using KML instead of a specially designed JSON file. First of all, KML is very powerful in terms of geocoding, having capabilities of representing complex shapes, view angles and geographical information. One of the fundamentally important benefit KML brings to our project is that it can represent polygons, which enables us to confine the point of interests, buildings, piazzas, canals, etc., precisely. Next, KML is a proliferated and well supported standard for representing geographical data, hence it is well supported by most of the important geographical information systems (GIS) such as Google Earth or ArcGIS. This compatibility makes the completed database independent of the visualization tool. The same KML data can be imported to GIS software and the data complying with the KML standards is displayed without any special programming being necessary.

Visualization of ArtMapping Venice KML in Google Earth

Visualization of ArtMapping Venice KML in Google Earth

Another benefit of this standardization is that there are online collaborative KML editing tools such as My Maps interface of Google Maps. The availability of a collaborative interface enables the project to be extended if necessary by enabling multiple users to edit content simultaneously. Furthermore,  using a known standard and having collaborative interface creates the possibility for other enthusiasts to create maps that can be displayed in our interface.

Ponte di Rialto using Google Maps and KML

Ponte di Rialto using Google Maps and KML

Need for accessing individual KML placemarks and geoxml3 processor.

As was stated in the previous paragraph, KML is one of the most widely used standards for representing geographical information, and GIS software such as Google Earth or ArcGIS have strong KML support. Although these two systems are very powerful in terms of geographical representation, we have decided on using Google Maps due to its convenience of use on the end-user side. Although KML data can be imported and displayed on the map by simple function calls using Google Maps JavaScript API v3, the programming interface does not give the ability to reach the individual KML placemarks, the single entries for geographical locations. Since we need features that are not in the Google Maps standard, such as interaction with a sidebar and specially designed modal windows for displaying historical paintings, we need to be able to reach each placemark individually. In order to do this we decided to use geoxml3, which is a KML processor designed for using with Google Maps. This KML processor satisfies our needs, since the KML attributes parsed by the code can be explicitly stated, hence we can have interaction between images and geographical locations.

Display of prints related to a location.

Display of prints of Piazza San Marco in a modal window.

Database of location information and the paintings of Venice.

Since we have decided on how to store the geographical information, one last decision was to choose how to store the information of paintings. Since the beginning of the project with the test JSON file, we have been thinking of representing the image and geographical information in the same file. However, since we have multiple paintings of a single location we needed to “hack” the KML standard to meet our needs. After some discussions, we decided to separate the database of the paintings from the geographical information and have an interface between them for interaction.

Splitting the geographical and painting information brings a number of benefits. First of all, with this decision, in the end of the project we will have one KML file that represents historical locations of Venice and an XML file documenting the paintings in the book Merveilleuse Venise. These two databases can then be used by themselves for different purposes, such as indexing paintings for the image database or visualizing the locations in a software for the other. Furthermore, it enables the users having historical paintings to add the information by dealing with only the image database and not going through the KML file.

Relief of finalizing decisions.

These two decisions finalizes the questions of how to represent and interact with the database. Since we have been coding the visualization interface while trying different representations of data, we have the necessary code for displaying information in a user friendly way. With the relief of finalizing the decisions, the only fundamental step remaining for the realization of ArtMapping Venice is to complete the database for the paintings and their associated geographical locations.

Choosing the Right Mapping Tool

Just after solving the issue about the content of our Venice art collection (please refer to A Collection of Ancient Prints of Venice for further information), our research started to appear much more complex than we have anticipated. The reason for this surprising difficulty basically stems from the fact that the amount of possibilities regarding the coding of information and their visualization on a map is numerous.

Even if, according to our previous plan, these weeks should be dedicated to the organization of the metadata, we believe that the choice of the system on which we should devise our ideas is the priority, and we decided to spend these days discussing about this issue.

Maps or Earth?

The first problem regards the choice between a system based on a flat map, like the one presented by Google Maps, and a more sophisticated approach that could make use of a more realistic 3D map, developing a platform based on Google Earth. This is a critical decision since on this will depend the software used for implementing the maps.

Revolutionize the conventional approach for the discovery of a city.

An “earth” representation would certainly be richer than the mere “map” visualization, and could be also more suggestive if applied in a project like ArtMapping Venice: exploring the city in three dimensions aided by Google Earth would be very well suited if our aim is to fascinate the users. Nevertheless, in our opinion ArtMaps are based on a slightly different concept: the ArtMap arises from the need to have convenient method to integrate the information one can find on a travel guide, with the final aim of revolutionizing the conventional approach for the discovery of a city.

Led by this idea, we do not want to develop a platform using Google Earth just because it is more complicated due to its capabilities; on the contrary the first aim of our project is to provide a stable and useful tool that allows convenient visualization of artworks in the right location and provide the basic information required. In conclusion, using Google Maps instead of Google Earth would certainly improve the portability of the platform, providing a simple and user friendly interface and avoiding the need for downloading plug-ins required for running Google Earth on a browser.

Choosing the right mapping tool in light of the design specifications

After deciding on using a “map” based platform, next decision we faced was choosing between using Google Maps or a more professional Geographical Information System such as ArcGIS. In order to have a fair comparison, we have built a sample page having representative landmarks with both tools. It was seen that creating landmarks and displaying the information visually is very easy for ArcGIS. The online interface enables the user to enter structures supported by KML graphically, thus saves the user from needing to write code. However, the information that can be entered through the user interface is limited: for example, the user interface by default gives the user opportunity to enter only one image attributed to the location. However, this by itself does not meet our design specifications since we want to have an interface enabling compare the past and present state of Venice. Furthermore, we want to design a time axis for the appearance of locations where we have multiple drawings from different years, which increases the number of images attributed to a single location and special interface design to accommodate a time slider.

These two factors, in our opinion, is more important for the sake of ArtMapping than having an easy to use interface to create the database. For this reason, we believe that the best choice for implementing ArtMapping Venice is the Google Maps platform.

Google Maps

Sample entry using Google Maps. Our present implementation allows usage of “pins” to mark point of interests.

ArcGIS

Sample entry using ArcGIS. The point of interest can be marked by any shape supported by KML, in this instance the marker is a polygon.

A Collection of Ancient Prints of Venice

We have just come back from our academic holidays…but what’s happening? It seems that Venice has become a very cool city among the digital humanists! In February, EPFL and Ca’ Foscari University founded a joint program and started “The Venice Time Machine” project that aims “reconstructing Venice’s past, better understanding its present and anticipating its future”. The program is highlighted on numerous media including the world famous magazine “Le Temps” and the EPFL homepage.

The Venice Time Machine

The Venice Time Machine on IC-EPFL Homepage

The team of ArtMapping Venice project is very excited to see that Venice is increasing its importance in digital humanities community, and comes back at work motivated by a new inspiration.

As for the ArtMapping Venice project, well, how could we ArtMap Venice without art? The goal of the first two or three weeks is indeed the collection of artworks relevant to the history of Venice. This is a very demanding and crucial step for the project, since the following work depends highly on how good our dataset captures the geo-temporal evolution of Venice. Unfortunately we can’t have hundreds of people available for working at the database, and it is of fundamental importance to understand where to get the pictures we need.

During the planning phase of the project two months ago, the web seemed to be our primary source, we have had time to meditate on this and to define some pros and cons. Obviously, the web offers extensive artwork resources, and it would be quite easy to find any artwork only typing the title or the artist in the Google search bar. Another interesting idea was exploiting the existing ArtMapping archives (for more information please refer to the previous post here). However, besides the problem of where to find the artworks, the issue of what to collect is much more complicated. An example is represented by Piazza San Marco, widely diffuse subject that has been portrayed several times by artists of different ages. Since it would be impossible to cover completely Venice’s art, we have to establish a research method: focusing on a particular category of artworks, the confusion would be avoided and there would be always the possibility to extend the database adding new items.

Merveilleuse Venice, a merveilleuse book.

The solution was clear when we had a look at a very particular book, “Merveilleuse Venise” by Sophie Monneret published in 1971. The book contains drawings of Venice dating from 15th century to 18th century, drawn by famous vedutisti including Canaletto and Michele Marieschi, that are preserved in the collections of Bibliothèque Nationale, Paris and Musées Civiques d’Art et d’Histoire de Venise. The drawings that are found in the book are of various nature including city plans, cityscapes (vedute), depiction of public events and daily life in Venice. This variety of nature of drawings is especially suitable for our project, since it will allow us to capture the historical Venice in a more thorough manner. Furthermore, the drawings are accompanied by explanatory text conjugating the drawing to history, which is again a big plus for conveying the story of Venice.

The advantages of collecting the prints from one single book will allow us to work in a systematic way, also making use of a reliable source of information. This is very important if we want to achieve a reliable temporal resolution in the ArtMaps. A nice example of how the time factor is important is shown in this print showing Campo di San Geremia by the Domenico Lovisa dating to 1717.

add caption!!

Campo di San Geremia, 1717

Past and Present.

For contrasting of the past and present, we used Google Maps to find the geographical location of the drawing and Wikimedia Commons to obtain the present appearance of the location. What we have found is really interesting: although Palazzo Labia (two story high building on the left side of the image) and the campanile looks almost the same the Church of San Geremia appearing in the centre of the drawing is very different than the one in the photograph. After a study of the history of the edifice, it was seen that the current church was built in 1753, three decades after the drawing. The team of ArtMapping Venice thinks that this is a very sound example of contrasting past and present of Venice.

San Geremia

San Geremia, 2011 – Wikimedia Commons

Apart from artwork collection, there are updates on the webpage for the project. We have set up a homepage that can interact with Google Maps through Google Maps API and displays pins on the map on locations obtained from a JSON database. When the pins are clicked, an information box opens up displaying information about the location and the drawing gathered from the same JSON database. A screenshot displaying the homepage can be seen in the figure below.

Homepage

Test homepage for ArtMapping Venice Project

The reason of choosing JSON for the initial testing of the database is first, JSON is extensible so we can add more data fields as we need as the project progress and second, Google Maps API has a very good interface for handling JSON files. Some drawbacks now with the current implementation of the database is that, the data is entered by opening the JSON file and typing the location, the name, the coordinates and such by hand. Although this is a valid approach, it is not feasible as the database grows so a user interface with preferable automated operations, such as obtaining the coordinates automatically, is more beneficial for the extension of the project. Another drawback is that the database is structured with a standard of our own defining and thus does not comply with well known markup languages for representing geographical information such as KML. The benefit of using such a standard language for representation is that the database would be independent of the way we display it, giving us the possibility to use more professional geographical information systems (GIS) such as ArcGIS.

Methodology

There are four main milestones in realizing the project. Although the complete implementation of the items listed below depends on the completeness of the ones listed above them, partial simultaneous work can be applied by team work. The milestones in realizing the project are as follows:

  • Data selection
  • Data organization
  • Building database
  • User Interface Design

Data Selection

Venice, playing a crucial role in Renaissance and being one of the essential destinations of Grand Tour [1], a 17th century tourism boom within aristocratic Englishmen, has extensively been the subject of paintings. Due to these two factors, it is relatively frequent that one finds the veduta, highly detailed painting of a city scene, of the same locations of Venice by different artists during various periods. In order to achieve the goals of the project, an artwork survey needs to be conducted in order to visualize the appearance of the scenery through the time axis.

For thorough artwork survey, it is beneficial to start looking at the collections of the online tools conjugating “history and art” and “geography and art”. Some examples to these tools can be seen in the state of art post in this blog. Although the context these websites present the data they have differs from the goal ArtMapping Venice, their database provides a good starting point in terms of obtaining knowledge about the important location of Venice, the recurrent artists and the period span of artwork.

Besides doing a survey of the databases of projects relating “history and art” and “geography and art”, the works of painters who are famous of their paintings of Venice are going to be surveyed. In this context it is important to note the following painters Canaletto (1697-1768), Michele Marieschi (1710-1743), Bernardo Bellotto (1721-1780), who are known to be the most famous vedutisti, veduta painters, of their periods. Furthermore, the portfolio of famous landscape painters, such as J. M. W. Turner (1775-1851), supply sketches besides paintings, which describes the city more exhaustively but with less details. However, these sketches can be used if the details are adequate.

Data organization

Having the same importance as choosing the data to be used is the representation of the metadata associated with it. The name of the painting and the artist are two fundamental information. However, to fulfill the goals of the project further information needs to be conveyed. Due to the goal of representing the location on time axis, the date and the location of the painting is crucial for our needs. The location of the painting contain the coordinates of the subject of painting since the location is going to be shown on the map, the name of the location which may be the name of a piazza, a building or a canal. Since the locations are going to be represented on time axis, paintings of the same location needs to be linked together. Furthermore, a scenery painting can contain a wide viewing angle which can contain more than one point of interest. Such as the painting of Piazza San Marco by Canaletto. In the painting there are the St. Mark’s Basilica, St. Mark’s Campanile, Procurati Veccie, Procurati Nuove, Clocktower and a part of Doge’s palace. All of these points of interests needs to be documented and made available for the search utility of the project.

piazza-san-marco-with-the-basilica-1730

Piazza San Marco by Canaletto – 1730

Building the Database

There are two main options of building the image database. One of them is downloading the images to the server, and the other one is using hyperlinks for the images. The first option is more robust so the links may change in the second case making it necessary to traverse the images to check if they are still visible periodically. However, the second case provides a broader data set so that it makes the use of data copyrighted against downloading.

For the metadata representation XML or JSON are two important options. Although it can be argued that XML is more popular and more proliferated than JSON, Google Maps recommends the usage of JSON over XML for geocoding [2]. The advantages of using these standards are the extensibility and human readability. Furthermore they both can be incorporated into AJAX, a group of web development techniques where updates are received from the server without changing the current display,  for a more fluid interface design.

User Interface Design

For the user interface implementation for the project three different models draw attention, each having different pros and cons.

First of the models is to develop a software that is not public. It connects to an online database to retrieve paintings during the application runtime. This model has the advantage of making available using a broader dataset since the tool is personal use. This implementation gives the possibility to use material that is made available for personal use only. Furthermore, since the software interface is going to be designed from scratch we have a broad possibility on the interface design.

Second and the third approaches are similar such that they are both browser based. One of them is building a public webpage that can be accessible world-wide and the other one is building a web tool but not making it public. Again, the second option has the benefit of making available of using a broader range of material while the public implementation narrows down the material that can be used due to copyrights. However, a publicly accessible webpage has the benefit of allowing crowd-sourcing which broadens the limits of what can be accomplished. Both of the approaches here have the advantage that there are well developed browser based map application programming interfaces, such as Google Maps JavaScript API, and style design tools, such as CSS, which eases the design and implementation.

Irrespective of the methods listed above, the fundamental wireframe of the design is as follows. The tool provides the end-user a user-friendly graphical interface displaying the map of Venice having pins on the locations that has artwork affiliated with them. On the screen, floating boxes gives the user the ability to narrow down the artwork to specific artist and period, and the utility of searching a location, painting or painter. As the user clicks on a pin, a pop-up window appears displaying the present appearance of the scene with the painting in a slip-screen fashion. The time slider on the bottom of the pop-up window can be used to display paintings of the same location during different periods when existent.

References

[1] The Grand Tour, Heilbrunn Timeline of Art History. Web. http://www.metmuseum.org/toah/hd/grtr/hd_grtr.htm. Date of access: 18 Dec. 2012

[2] The Google Geocoding API. Web. https://developers.google.com/maps/documentation/geocoding/. Date of access: 18 Dec. 2012