Companies from all sectors are looking for the best way to achieve their Sustainable Development Goals and explore how their actions impact the United Nations 2030 agenda. However, it is not always easy to measure the impact of the actions carried out at a corporate level, or the communications that are made  oto form opinions from the outside on the development of these objectives.

For this reason, at VASS we have worked on the development of a tool based on knowledge graphs, built  with technology so that companies can achieve their sustainability objectives.

A knowledge graph tool to understand the conversation about the SDGs

VASS have developed a new tool for one of our clients based on knowledge graph technology, capable of measuring engagement with the SDGs. This opens the door to finding new business models that take into account the goals, bringing us closer to sustainability.

With this tool, it will be possible to identify which SDGs are being talked about through different communication channels. It will also help to understand to what extent or with what sentiment they are discussed in the societal debate and how engaged customers are with different SDGs.

How does the knowledge graph work to identify this information on the SDGs?

In the specific case of this project, developed for a financial institution, we started by extracting information from the UN website. Thus, from the information on each SDG, a division is made into phrases and in turned into words, all with the aim that the graph is able to classify the news about the SDGs.

A word disambiguation process has also been carried out to find out which goal is being talked about and to what extent, detecting words with more weight in the text. Similarly, a sentiment analysis is carried out to differentiate between positive, negative and neutral.

How do knowledge graphs work and how do they differ from AI?

A knowledge graph is actually a database. The difference is that traditional databases are data oriented, while networks are more oriented towards the relationship between the data. The focus is therefore on the way knowledge is stored in the form of a graph, with connections between information. The query language of knowledge graphs is designed to take advantage of these relationships, and even find indirect second- or third-degree relationships.

One advantage that graphs have over Artificial Intelligence systems is that Artificial Intelligence needs continuous training of the models, including new information. However, knowledge graphs are able to find complex relationships without this constant training.

What other applications can a knowledge graph tool have?

Beyond this specific example, knowledge graph technology can be used in different sectors and use cases, replacing traditional classification systems using Artificial Intelligence. For example, by finding indirect relationships between data, it could be used in the financial sector to detect a fraudulent transaction. It will also serve to identify what is being talked about in social networks, news, etc., whether it is about sustainability actions or any other topic that we can configure.

 VASS is always in a process of continuous innovation, especially when we able to us simplify and provide real solutions. That is why we continue to research on the applicability of knowledge graphs, the possibilities of combining them with Artificial Intelligence or even the ability to improve efficiency at an operational level.