Objectives and challenges

Despite the fact that art and culture play an important role in our lives and the actual size of the industry surrounding them is massive, there is an objective difficulty in accurately measuring the actual impact of culture and art related events or projects. Although some notion of success can be quantified through very basic metrics such as the number of attenders in a concert or the number of paid downloads or views, there are many cases in which no economic transaction takes place, and, most importantly, in all cases there is no way to measure success in terms of impact or influence. The existence of a clearly measurable quantitative metric would be able to guide the cultural industry into more informed decisions which would result in an upgrade of the general cultural product.

The main objective of this research is to provide a solution on how to measure the impact and importance of audiovisual art projects. Since media is the main institution which accounts for the formulation of public opinion and trends in all major aspects of social life including art and entertainment, it makes sense to consider media coverage as an important indicator of influence. Studying and analyzing all aspects of media coverage surrounding a cultural event can provide a notion of how big its outreach was and to what extend it achieved success in terms of influence and social impact. But this sort of analysis is both subjective and time consuming if approached through traditional means.

In the modern world, information about everything can be found online on the World Wide Web. The abundance of data publicly and readily available on the Web creates an opportunity to measure the various parameters that define the outreach of an audiovisual art project and make it achievable to quantify its influence. The use of computational engineering tools to process large amounts of data extracted in an automatic way from the Web by means of scraping techniques can provide the industry with a new method by which to measure impact. A data-driven research approach to this new information will lead to the generation of multiple hypotheses that can be tested through statistical analysis and can lead the industry to new insights.

A major methodological issue in the field of communication and media studies in the past has been achieving an acceptable sampling size that would allow the use of quantitative methods. In the recent years and given the continuous increase of relevant data availability, the main problem has shifted to defining the appropriate metrics that would lead to the discovery of meaningful patterns within large-scale datasets. In addition to that, many approaches and research methodologies tend to be either over-simplistic or over-complicated, so this question of balance must be addressed. It is imperative for a methodology to reach an understanding of how to deal with the quantitative notion of impact.

To that extend, this research aims at the development of a theoretical model which by taking into account all the major characteristics of audiovisual art projects and through measuring their outreach on the Web can provide numeric indicators for their influence that will allow for comparisons between them. The key aspects that will govern this model are the use of specific indices and metrics for the different creative and productive characteristics of each project as well as for its media coverage outreach, the need for minimal human evaluation and intervention that will aim to deal with the qualitative aspects of the research material both in identifying characteristics and in measuring outreach and an overall optimized level of complexity.

Having established such a theoretical model, this research proceeds with its implementation through a fully functional repository of audiovisual art projects that take the form of an online Web application. This repository is being populated with information concerning both the characteristics and the outreach of any audiovisual art project that is part of the general cultural landscape of Greece. This system will not only serve as a proof-of-concept for the capabilities of the theoretical model, but also at the same time provide a much-needed observatory of the vast number of cultural activities in Greece and their potential impact. In order to achieve maximum transparency, all parts of the abstract model (e.g. “art project”, “Greek project”) is clearly defined and placed into boundaries. This is an essential part of the implementation and helps avoid an excessive broadening of the purpose and use of the model. Taking it a step further, the Web application also provides a RESTful API which allows the scientific community to access the repository and use the information available to their own ends.

Scientific and/or social impact

During the last decade, information technologies have generated an enormous amount of data and its accessibility has given researchers a new way of producing data-driven scientific knowledge. Cultural institutions, based on openness and transparency, have adopted principles that contribute not only to the creation of cultural  items on a digital media format but also the promotion of non-digital items through the use of the digital space. 

The large collections of such media combined with the trend of sharing them through the virtual communities created on the World Wide Web and Social Media platforms offer great possibilities for the study and social  analysis of culture and more specifically audio and visual art, through the means of big data and computational  engineering. Cultural research based on big data, benefits not only from the range and variety of the available  information which facilitates a quantitative approach and the detection of behavioral patterns and emerging  trends, but also from the significantly reduced time and resource requirements. The proposed research, through its data-driven approach, contributes to the notion that studies of audiovisual  arts can be greatly benefited by complementing their research methodologies with practices that are based on  the analysis of big data collections related to culture as well as the analysis of Web resources like websites and  Social Media posts. Furthermore, it reinforces the standpoint that big data driven analysis is most suitable for  identifying and, most importantly, predicting relationships, patterns and trends in the wider field of arts and  culture.

Additionally, the process of data collection for our analysis produces two research tools that could benefit future studies. Firstly, an important product of this process is a well-documented and tested research method and implementation for data extraction regarding audiovisual art projects from Web resources like websites and Social Media, that even though is implemented for a strictly designated region, it can easily be expanded to facilitate similar research worldwide. Secondly, the data gathered through data extracting and content analysis, which concern the Greek audiovisual art landscape, is organized into a data model that  makes use of Semantic Web technologies and the metadata regarding these projects and events are presented as Linked Open Data and are readily accessible from researchers to be used in further studies  through a simple, lightweight and fast RESTful API. Going beyond its research facilitating qualities, this collected  and organized information can also be used by artists, creators and organizers of audiovisual art projects and events as a means to evaluate the potential impact of such an endeavor project and more importantly to guide  decision making for future attempts in this field, thus helping raise the overall outreach performance of such  projects. 

The most important contribution of this research to the scientific community is a better understanding  regarding the correlation between the creative and productive characteristics of an audiovisual art project and  its achieved outreach. The results of this research provide insights with respect to how each parameter (type, size, promotional funding, target audience etc.) related to an audiovisual art project can benefit or hinder its  presence in Web based media and hence it’s overall outreach. Through this process we expect for patterns regarding the organizing of cultural events along with trends for future projections to emerge. Furthermore, the results have a social impact, as they provide artists and cultural event organizers with a guide regarding the impact of each creative or production characteristic of an art project or event on its  achieved outreach and media coverage.

Last but not least, for the main purpose of this research, a model for the observation of the correlation trends between the characteristics of audiovisual art projects and their outreach are created and an observatory which monitors these trends has been established. This model is distinctly defined, described in detail and comprised of multiple indices of quantifiable information. Although our study focuses on projects and  events which are part of the cultural landscape of Greece, the proposed model used has become available to  researchers as a tool for related studies that focus on other regions or target a specific demographic group. Additionally, through the results of this research and the implementation of the proposed model, a valuable practical tool is available to the cultural industry to foster informed decisions regarding the evaluation and public or private funding of art projects.

Research Methodology

The process of analyzing the relationship between the creative and production characteristics of Greek audiovisual art projects and their achieved outreach begins by identifying the metrics and defining the indices  needed to evaluate the multiple aspects of outreach they might have. Using this knowledge, appropriate data structures have been developed and used in order to store and serve all relevant information through an online  repository. This repository has been populated by processing the information gathered from the World Wide Web and Social Media through data extraction and converting it to the appropriate structured data format using data manipulation algorithms. Additionally, the processed data is validated and complemented through human  interference by means of content analysis. Over the course of multiple months, the data extraction process is continued with its focus shifted towards gathering information about the outreach of each event in the repository.  The sum of information collected and the application of a clearly defined theoretical model to this information is the primary role of the data-driven observatory of Greek audiovisual art projects on the Web. Monitoring  how these overall metrics and indices evolve over time and by use of statistical analysis, multiple hypotheses have been generated and useful conclusions are drawn. A comprehensive graphic presenting the various assets that comprise the main mechanism of our research methodology can be seen in Figure 1.


Figure 1 

The first step towards achieving the purpose of this research is defining a model for the quantification of both  the characteristics and the outreach of a specific audiovisual art event or project. To some extent, the  particularity of the audio and visual arts landscape of Greece has been taken into account during the creation of  this model. An initial classification of the creative characteristics of each project is essential. Based on the form  of art used and factors such as techniques, materials, the artists involved etc. projects and events have been organized through a multilayered category structure. That made it possible to study outreach with regards to  the nature of the creative work. Additionally, the different production characteristics of each project must be  identified and measured. Attributes such as the scale of a project in terms of production value or in terms of  creators, participants, potential visitors or viewers, the public funding of a project, its locality, its target  demographic etc. can be converted into individual metrics. Similarly, the indicators that can be used to quantify outreach are identified. These include relevant online media coverage not only in scale but also in  intensity and popularity, Social Media interest both over the course of time and at its peak, target audience range etc. The sum of outreach information is quantified and combined to form an elaborate scale that can be  used to measure the potential impact of a project. Finally, the relation between different metrics are being expressed through the creation of specific indices. Both the process of classification and of identifying metrics  and defining indices are driven by the expertise of the researchers in the fields of media and the audio and  visual arts, as well as the extended study of all relevant bibliography with additional focus to the latest  developments in the wider scientific area and the cultural industry.  Having established all the attributes and characteristics that outline the nature of audiovisual projects as well as  their outreach, the task of creating an appropriate data model to encapsulate this information has begun. It is  essential for the purposes of this research that all information stored and served by the repository adhere  to the principles of the Semantic Web. Towards that end, the Dublin Core schema of metadata is used as a basis for creating the ontology that governs our approach and special effort has been put into using Linked Open  Data as the values of the different terms of this ontology. The main concept of the ontology is the project  itself. Its terms comprise mainly of the classification qualities, metrics and attributes identified during the aforementioned study process. Special emphasis has been given to the spatial and temporal properties of each item. In an effort to accommodate the use of LOD, art and culture vocabularies will be used, while the spatial  properties can be described through the use of Geolocation URIs. In order to serve the purpose of this research  better, a separate concept has been used to describe the various media sources from which information relating to each project or event is derived from. Information relating to each media source might include attributes such  as the type of the source, its popularity, its reach, its target demographic and other similar aspects of its nature  that can be used to infer the equivalent qualities of the project or event that the media source relates to at any  time. The process of evaluating the presence of a project in a specific media source has been multifactored and  play an important part of inferring a project’s potential impact. During the process of designing the ontology, other potential concepts has been considered such as a concept that might hold information regarding artists or  creators or a concept that might describe specific representations of a project or event on the World Wide Web.