Interview with Gábor Bundschuh, Head of Development & Innovation at D-TAG Analytics Inc, a media intelligence company in the US.
Hi Gábor, what is your background and what is included in your current role at D-TAG Analytics?
I have been working with different Information retrieval related solutions for more than 30 years and I have relatively strong experience in AI/ML/NLP tools, data and information management.
D-TAG’s main goal is to support the decision making challenges of companies as effectively as possible. We have a self-developed solution for analysing any type of unstructured, textual contents, especially social media posts and/or documents.
In what ways does D-TAG Analytics differ from other companies?
At D-TAG we are trying to identify the special demands of our potential customers in industries we consider very important, such as pharma, insurance, and banking. Our solution covers both legacy type data sources and social media; it is a hybrid solution for heterogeneous data.
With the help of a well-configured and sophisticated data pipeline and fine-tuned NLP background processes, we enrich the data to be analysed effectively, and turn them into information.
The final goal is to have information instead of bits and bytes.
What are your greatest challenges ahead at D-TAG Analytics, when it comes to serving your customers and developing your offering?
The fine tuning of data (effective data enrichment and metadata handling) is a particular challenge because the quality of data is the most important prerequisite of the analytics. The clarity of the collected data and the quality of the data (metadata) enrichment process are the key points of our approach. Based on data with good enough metadata backgrounds, we can use efficient taxonomies, topic management, entity extractions, and other means in order to be able to discover a large amount of information and — very importantly — to give further ideas to the customers about what they are searching for or what they can be interested in. With the help of this information, we cannot only answer predefined questions, but we can also use the results to formulate new questions and get answers to them very quickly and efficiently.
All customers come with a different level of knowledge. What is important when taking on a new customer?
The starting point in a project is standing on the same page with our customer. At the same time, we want to give them as much experience and knowledge we have as possible, and explain how a typical information retrieval or analytics solution works, what the information is trying to show, and how they can understand it. I think that beyond the predefined business requirements, it is also important to give new ideas based on some “hidden information” we were able to discover.
When it comes to the actual data behind the media intelligence you do, what kind of data or media not currently used can be interesting in the future?
Most solutions on the market are able to handle any kind of data format coming from any kind of repositories, on premise or cloud based. On the one hand, handling the rich media content (picture, voice, video) effectively is still a challenge, especially in the case of special languages. We are continuously trying to find ideas and tricks in order to improve the quality of the speech to text processes. On the other hand, it is very important to improve the quality of AI/NLP/ML related processes, because as I have already mentioned this can be the token of the excellent results. The level of a successful automation will depend on the granularity and complexity of business demands.
Can the entire process of media intelligence be automated in the future?
I think that the short answer is yes, but the detailed, longer one is no. The basic elements and the processes around important milestones of the information management pipeline can be almost fully automated, the level of the automatisation will increase, but the detailed nuances will still play important roles, and will use ML capabilities. But they require manual corrections, improvements and considerations as well.
How do you think the media intelligence industry will change in the next 5-10 years, and what are the greatest challenges and excitements ahead?
The quality of the analytics will increase, the range and type of customers and end users will expand, and at the same time, the tasks will be more varied.
Changing customer needs pose an increasing challenge to us which won’t be easy to handle from professional nor from other points of view, although technology will also continue to evolve in the meantime as well.
Understanding the relation between the existing data contents and the customer’s business needs will remain one of the most important tokens of a successful project. Technology will help to understand this relationship more deeply.
By Peter Appleby