Thanks. Hello. So my name is Vasily Gikazyaki and I'm a data engineer with Wikirate International.
And I'm going to talk about how Wikirate empowers communities for transparent corporate impact
research and analysis. But before we get into details about what Wikirate is, I would like
to talk a little bit about what is the problem with environmental, social and governance data of companies.
So usually when it comes to ESG data, we can say that they are expensive, exclusive and inconsistent.
There are a lot of data sets hidden behind paywalls. So individuals need to pay thousands of euros per year to get access.
Additionally, there are a lot of ratings. There are a lot of organizations actually producing ratings about companies.
But the problem is they don't provide access to the low level data sets. So it's difficult really to understand what they are rating.
And also, yeah, they don't make the methodologies transparent or the sources transparent.
Yeah, finally, the last few years companies started reporting more ESG data in a text format, in sustainability reports.
But the problem is that a lot of company reporting is not standardized and that hinders large scale analysis and comparisons between companies.
So what makes open research so important in the context of corporate accountability?
It actually fosters transparency in corporate practices and empowers different stakeholders, especially people that don't traditionally have access to those data.
That they don't, let's say, investors and they don't have the money to pay to get access of this ESG data.
It encourages collaborations in global scale and promotes data driven decision and policy maker making and drives positive changes.
So Wikiread is an open source, open data platform that brings corporate ESG data together in one place, making it accessible, comparable and free for all.
It's a wiki that means that anyone who has a passion for sustainability and ESG data can come to the platform, contribute to the research, contribute to the available data and organize their research as well.
So our community is mainly comprised of civil society organizations, academics, university students, data and sustainability enthusiasts.
And we strongly believe that in research and in research in companies everything starts with a question and ends with an answer.
So I would like to give you a sort of overview of the structure of the data on Wikiread.
So these research questions we call them metrics and we can have for its metrics several answers.
And its answer is linked to a specific company, a year of reference and a specific source.
So here we have an example about, did Airbnb UK Limited produce a statement in relation to any modern slavery legislation or act in 2022?
And the answer in this case was yes.
It produced a modern slavery statement under the UK Modern Slavery Act and there is a source and citation actually linked to this answer that leads to the actual modern slavery statement of a specific company.
So in Wikiread in addition to research metrics we provide also calculated metrics as a tools for calculations and for analysis.
We can say that the research metrics are actually the building blocks for analysis and the calculated metrics actually are built on top of research metrics and allow users to run calculations.
So we do have namely Wikiread in score metrics, formula metrics and in more strictly formula metrics allow users to run their own calculations in coffee script so they can be quite complex or not complex.
It depends on what the users want to do with the data.
So and this actually, this calculated metrics helps to bring transparency into ratings.
Here we have an example with the fashion transparency index is actually a rating that scores companies, fashion companies based on how transparent they are on different sustainability topics.
We are a partner, we have a partnership with Fashion Revolution which is also an NGO and they're building, they do this analysis in research and we're helping them to actually make the research, the data, the ratings, the analysis, everything transparent available to the public.
So one source of data on Wikiread is of course data that are coming from the ground, from civil society organizations but also there are a lot of data in the public domain.
So it's easier to bring structured data and semi-structured data and building data integration pipelines to Wikiread but of course we do have the challenge of unstructured data and how we are going to bring those kind of answers into the platform.
So for those reasons we are performing research projects and we have called for volunteers to come and research those reports and finding answers to questions on specific topics like modern slavery, greenhouse gas emissions, etc.
So how is the data used? One use case of Wikiread data is building data dashboards that are actually used for advocating for change.
One example is Fashion Checker which was developed in partnership with Klinglo's campaign and actually advocates for worker rights, especially worker rights on the supply chains of fast fashion companies.
We have the Beyond Compliance dashboard that was also a partnership with Work Free Foundation and is a living data dashboard that assesses modern slavery reporting and tries to highlight gaps on modern slavery reporting and puts for new legislations and for new policies.
Also the data is used for writing news articles and producing help CSO, CISO-Cyber-Suside Organization, produce reports and making research findings and analysis transparent.
So it's also used for writing research papers, research some of those and Wikiread data are free, are under Creative Commons license so it's welcome to anyone to use the data, explore the data.
They can do it through the API and through the user interface. We have an available RESTful API and several wrappers that will allow users to pull data from the platform, also contribute data to the platform if they want to.
We also have a GraphQL endpoint that allows users to form more dynamic queries based on their needs.
So where to start with Wikiread? If you're interested in contributing data, you can always say start with the guides please.
Read the guides as most of the questions are answered there.
But of course if you have any more questions you can directly contact us.
And yeah we have several projects that are in need of contributors. You can help us improve the data.
We have verification tasks, we have verification tasks so we ask the help of the community to help us with this process.
And yeah if you are interested in volunteering these links are available to the slides.
And yeah you can contact us if you want to share ideas with us, form partnerships or get support.
Yeah as I said in the beginning Wikiread is an open source project written in Ruby.
You can check out our github repository and if you want to get started with Wikiread and DECO you can do it.
And you can also create if you want your own data dashboards if you're interested in ESG data.
So yeah I think that's all from my side, maybe it was too fast.
But yeah thank you.
Thank you.
We have maybe four minutes of questions.
Hello.
Hi.
I have a question about if AI has helped you in a way in any of these processes for example while manipulating or getting data from the public domain or something like that.
Yeah so yeah we are considering, sorry the question was about if AI helped us in any way obtaining data from the public domain.
So the answer is that we are considering using now AI and LLMs for extracting actually more structured answers from text reports.
But still we are in the test process so yeah hope that I answered your question.
Yeah.
How much of companies are covered by the data set?
Is there specific industries you are targeting?
Yeah so yeah we have about, I'm sorry yeah.
The question was how many companies we cover at the moment on the platform.
So we cover about 1400,000 companies.
The biggest focus on research or the biggest companies.
So more data can be found at a very popular let's say companies and because we did have a lot of projects on the fashion industry we do have a lot of data about fashion companies.
And in total at the moment we have five million answers.
So yeah.
Any other questions?
Yeah sorry.
So it's open to contribution as far as I recall and you mentioned some verifiers.
I want to ask how do you make sure the data is consistent and how do you go through the checks and see if the data is reliable?
Yeah so yeah the question is how we check that the data that they are coming into the platform is reliable.
And of course it's a question also because we are doing crowd research sometimes people do not have the expertise on ESD topics.
And what we do is we have different verification levels.
So we consider an answer verified when more than two people have come to the same conclusion.
And you can see on the platform that we have Stuart verify and community verified.
Stuart's usually are members of the community that they have more expertise on the specific topics that the research is about.
I'm always the person that's like let's squeeze as many questions as possible so we'll do really rapid questions right now.
Very quickly.
I was just wondering if this could be expanded to cover other types of data rather than ESD.
The question is if this could be expanded to cover other types of data.
Yes it can.
It's again about environmental data but one use case that it comes on the top of my mind now we are having companies you can have something similar for countries.
So you can highlight for instance the electricity or water usage or CO2 emissions per country and not focusing specific on companies.
Thank you so much.