There is a need of companies, countries, and organizations to improve Research and Innovation policies. Most approaches rely on methods that are not reusable, applicable to other fields and produce hard to prove results.
“Helping decision makers understand highly interconnected technological landscapes and as a result make better R&I decisions using knowledge data. For this purpose biofuels will be used as testing ground. “
Based on this research goal, several research questions were developed. These research questions offer a more concrete and quantifiable way of evaluating the research and more particularly, how these stakeholders would be helped. On section 2.1, the reasoning and the more practical aspects of these research questions will be described.
The research is structured following the scale of the examined levels of analysis, from macro to micro. The first level shows the dynamic analysis of the studied system as a whole (worldwide biofuel research and innovation), and serves as an introduction to the complex system model. The second level of analysis focuses on a meso level focusing on the country level. After the meso level, the study narrows with the analysis of the system at the micro level, this corresponds to organizations such as universities and businesses. Finally, two complementary analyses will be made. The first focuses on the intricacies of the system itself and the elements that compose it - patents and publication in this case - . The second is a validation exercise where, as a baseline, a more basic approach will be used and compared the general approach taken throughout the thesis.
The research questions follow:
How does the research of different terms develop over time?
Differences in research:
How to characterize the similarity between two periods in time?
What differences exist when comparing different years in scientific research?
Are all years linearly related or do research gaps exist?
Is biofuel research correlated with the price of oil? What exact terms are the most related to it?
Is the price of a consumer good related with the research volume? What terms are most affected by it?
How to characterize the scientific capability of a country?
How are countries related from a capability perspective?
Do international capability clusters exist?
Does the amount of money of you have (GDP per capita) determine the space of possibilities or your technological freedom?
Does the GDP per capita play a part in the capability similarity of countries?
Is capability similarity related to international collaboration?
Differences in research:
How to characterize the similarity between two countries?
Is there value in characterizing a country’s capability as a spectrum?
How to characterize the capability of a given organization?
Do organizational clusters exist?
How to compare two organizations from a capability perspective?
Do universities collaborate more with universities or businesses? Do businesses collaborate more with universities?
Is there value in characterizing an organization’s capability as a spectrum?
How does the volume of patenting and publishing of a certain scientific term evolve over time?
Is there a bias towards patenting or publishing? An analysis of feedstocks, processing technologies and outputs.
Can we apply unsupervised learning techniques to visualize scientific clusters?
Term frequency vs term pair frequency? What characterizes better the scientific capability of a country or organization?
As a guiding approach to these research problems, the Design Research Methodology was adopted. The figure below provides an outline of the elements that characterize this method: which include criteria definition, descriptive study 1, prescriptive study, and descriptive study 2. These correspond to criteria definition, literature analysis, method development, and application/success evaluation.
Although the DRM approach makes for the richest analysis possible, this framework was used as more of a guideline than a step by step methodology. In fact, some elements of the framework require extensive validation and cycling throughout all of the levels that compose it. For this reason, the thesis focuses on two main elements of the DRM.
The first element used is the descriptive study 1, where, taking as a starting point the criteria (e.g. research questions), the analysis and pattern detection in the dataset was made. The second element used, to a less extensive extent, is the prescriptive study 2b where using the experience and some basic assumptions about the behavior of the research landscape, conclusions were reached. It should be pointed out that throughout the analysis; an effort was made to continuously validate the results, through investigation of underlying behavior.
Finally, as this thesis seeks to assist a group of stakeholders in making better decisions through data, these individuals, utilizing their experience and superior domain knowledge, can draw on their own conclusions using the data provided, which by itself leads to a type of prescriptive study.