Bisnode is a leading data and analytics provider in Europe with 2,500 employees and 150,000 customers across 18 countries. The document discusses Bisnode's use of context mining and data science techniques like machine learning, network science, and massive-scale analytics to automatically detect business events and extract data components like website finders, contact details, and bitcoin addresses from over 4 billion online documents to build scalable business information products.
15. Bitcoin Addresses
Combining
Bitcoin & Big Data
for Risk & Compliance
4B online documents processed to
analyze Bitcoin addresses and
transactions
D A T A C O M P O N E N T
16. Bisnode Big Data & Analytics uses Data Science
to create data components as building blocks
for scalable Business Information products.
Hello everyone
It is a pleasure to be here as representative of the Bisnode Group, who has started a couple of years ago a deep strategic transformation, showing now tremendous effects throughout all covered countries in Europe. The ambition is to move from an uncontested leadership to another – from a leading data and information supplier to the most wanted big data and analytics provider, globally.
This sounds like a big deal indeed, but we should put that in perspective. Remember that our industry started back in the middle of the 19th century, with…
…pigeons.
The original Bisnode was founded in 1859 – 158 years ago. At that time, the German-born Paul Reuter used pigeons to fly stock prices between Aachen and Brussels, with a proud fleet of 200 birds, and it operated for one year before the telegraph cables was put in service. At the time, it was hailed for its speed, accuracy, integrity, and impartiablity.
Fifteen decades later, Bisnode has become the leading European Data & Analytics company, with operations in 18 countries and 2,500 employees, helping companies find and manage their customers throughout the whole customer lifecycle.
We do this by pioneering Smart Data to enable our customers to make Smart Decisions, we can match and analyze our customers data with our data and the data generated by the connected world today - the so-called Big Data.
We have more than 150,000 customers benefiting from our Smart Data as a service. We help them manage their data, their risk and compliance, lead generation, and ultimately their revenues by reinforcing their overall decision ability.
We have an unique database of companies, individuals, real estate and vehicle data in Europe.
On top of that, Bisnode is also the largest strategic partner of Dun & Bradstreet, the global provider of business information for more than 250 million companies in 220 countries.
A big deal indeed.
But coming back to… today. Who am I?
I am quite new in the Bisnode organisation, for a good reason – I am a tech entrepreneur, and I was co-founder and CEO of Swan Insights, a Brussels-based big data company, acquired in January by the Bisnode Group. This acquisition was made with the strategic objective of supporting the transformation of the group from a leadership to another, as I said.
What was Swan Insights doing?
We knew that human-generated data was exploding since a couple of years, actually three times faster than structured data, on social networks, blogs and communities, and we also noticed that companies were not equipped to grasp the value out of this data – this external data, out of companies’ walls.
This was the problem indeed: this New data was unexploited. There is a goldmine in RSS feeds, in corporate websites, in local news, in social networks, in Open Data, in countless datasets. Companies and organizations leave digital footprints everywhere, that could be used in Business Information offerings, to improve companies’ DNA, to predict moves, to match business interests, to turn fixed data dynamic, and in real-time.
Existing Business Information offerings still relied on fixed, financial data, often dated from the latest fiscal exercise. But companies are made of everyday decisions, daily moves, actual business activity from professionals. They can’t be summarized in financial statements.
As we say, data is the new oil. External data is the refined oil that companies were missing, to see their future, to take more informed decisions, to generate qualified leads, to reduce churn, to improve business performance as a whole.
So, how can we exploit this new oil with business information? We developed for three years a big data technology collecting and combining this data with corporate data, in order to bring context to information, because context is the key to turn information into insights.
Our mission was to bring context in real time, by building links between disparate information sources.
But let me tell you the story of…
… Sofyan.
Sofyan is a young 21-year old software engineer, from Brussels suburbs, just fresh out of school in computer engineering (he learnt web development languages), and he was looking for an internship, hoping to work for a SME somewhere in Brussels after that.
We met Sofyan at Swan Insights, we hired him for an intership of 3 months, then we hired him.
He quickly integrated the company, created a football team with some colleagues, and is a really talented and smiling guy.
A couple months later, in January 2017, our startup is acquired by the Bisnode group, and is now part of an international team regularly flying from New Jersey to Chinnai, India.
Just imagine the change, the sudden boost that this young guy has lived? He’s now part of the worldwide technology and data elite, taking part in the evergoing revolution about big data, data science and artificial intelligence.
This is precisely that kind of encounters that make a corporate evolve, and Sofyan, just like the whole Swan Insights team, now part of the Bisnode family, brings a dynamic speedboat to that transformation.
And then, we got acquired and the best of both worlds combined. (clic) On the one hand, you have the leading European data and analytics group, partner of Dun&Bradstreet, providing the biggest business information dataset ever, and on the other side, you have a small team of PhD-level data scientists, with a big data technology, ready to be plugged onto this data ocean.
Together we elaborated what we call “Context Mining”, which is a Contextual Intelligence engine, whose objective is to turn information into actionable insights. This technology is now fully available in the Bisnode group for the 150.000 customers willing to go a step further with data analytics.
How does this work?
We first collect 4 main catagories of data sources: worldwide news from global to local feeds, social data (from generic and specialized social networks, including profiles, bios, publications, and connections – who’s connected to whom), open data (which is huge, since you know where to look, you can find anything), and negotiated or purchased data from suppliers, authorities, regulators or organisations).
And then we apply the right data science algorithms on that data, in order to create data components – data components are small pieces of advanced technology that bring magic onto data-driven business goals – and that you can combine to create packaged products that fulfil a specific business need or function.
I’ll show you some examples in a minute.
But to give you an idea of the technologies that Bisnode is now mastering, it ranges from machine learning, deep learning, network science, blockchain and of course contextual intelligence, that mobilises a bunch of advanced scientific methods. These capabilities are spread around the group to scale them globally.
An example of what is this value chain capable of is the recent new Bisnode product called…
Bisnode Business Events, directly released from the Swan technology, and magnified thanks to the Bisnode global footprint, the massive resources and assets.
This product detects all relevant business events about all companies worldwide from any relevant source from the internet, in any language, dan delivered in real time with a context, as a high quality material for Strategic Monitoring, Risk & Compliance, Lead Generation and other purposes. A business event is any fact happening in a company’s life, such as a new product, a new investment, a change in management, a stock variation, a M&A operation – more than 40 event types.
The product monitors daily
30M active companies
3M decision makers and business contacts
100k news sites including local sources
90k publications on social networks (daily)
Of course the challenge is to understand this massive load of information coming every day, so data science like natural language processing is key to achieve this.
Bisnode Business Events is a tool that bring contextual intelligence to business information, merging real time, scattered sources, relevant environment and actionability. It is the result of a consistent Context Mining approach to meet concrete business needs.
And as I said, this packaged product, Bisnode Business Events, is the result of the combination of different data components, themselves resulting from the combination of data sources and the appropriate data science.
And those data components can be used for other purposes, other products, or in specific projects, or even as such to complete specific tasks.
And this is Context Mining.
Let me give you a couple of examples of data components.
The Website Finder data component is an algorithm able to detect automatically corporate websites from a list of company names, even if the website address does not match the company name. When your company is called Bisnode, it is easy to assume that bisnode.com is the right corporate website.
But you have lots of exceptions, for instance the Belgian media group Cobelfra has its website on rossel.be.
So the algorithm analyses the website content to validate the match between the company and the website itself. We run this tool on databases of millions of companies, and we reach 97% of accuracy in the results.
Another example is the contact details. This algorithm, starting from a company name, will find any business contacts linked to that company, using different sources, including the corporate website contact page, but also professional social networks. This is very useful to update CRM databases for instance.
A third example is Bitcoin Addresses. Our NLP technology processed 4 billion online documents to analyse bitcoin addresses and transactions, in order to link those addresses to the corresponding companies. This component is clearly more from the R&D side, and it is also an objective of this process, starting from the data, to investigate new trends and their potential impact on our industry.
So to conclude, the journey from the data and information leadership, to the big data and analytics leadership, is built from the ground up, starting from the data sources, raw material of everything, new business oil, on which we apply the most recent findings in data science – this is why we closely collaborate with universities on that aspect – and with this combination, we constantly develop new ways of revealing hidden insights from the data, because contextualize information turns it into concrete, actionable insights.