StockPulse analyzes communication on financial topics in publicly accessible social media and news sources
Please introduce yourself and your startup StockPulse to our readers!
My name is Stefan Nann. I have founded StockPulse in 2011 together with my colleague from University Jonas Krauss after finishing our studies in Business Information Systems at University of Cologne and Massachusetts Institute of Technology (MIT). At University and later at MIT we were involved in academic work. For example, we did a study in the field of collective intelligence where we anaylzed communication in movie forums to predict the Academy Awards. We also started to research about the correlation of Social Media discussions and the stock market during this time.
StockPulse analyzes communication on financial topics in publicly accessible social media and news sources. Software evaluates hundreds of thousands of opinions and news articles every day and shows at a glance, how much and in what kind of mood people discuss about financial markets. Using latest Big Data technology, communication on stocks, indices, commodities, currency pairs and major market events is monitored and evaluated in real-time 24/7. Moreover, the software automatically generates buy and sell signals for different asset classes. By now, our database contains historical data of more than six years, enabling comprehensive backtestings. The software and trading models were reviewed and confirmed by independent scientific studies.
Web crawlers are constantly searching and scraping hundreds of different internet sources for relevant financial topics and communication. Crawlers collect tweets, message board posts, news articles, and comments to news articles. Further, we collect and maintain all relevant persons/insiders of stock market listed companies and how they are related in the news. Text analysis works for English, German, and Cinese. We have a specific and deep coverage of the German speaking community.
How did you get the idea to StockPulse?
Jonas and I attended a class at the end of our studies at University which was about collective intelligence and the wisdom of the crowd. Our task was to crawl a forum on imdb.com (“Oscar Buzz”) and derive insights based on the data we have collected. We created a ranking of Top 20 movies in January 2007. The decision which movie received an Oscar was made in February 2007. We had a hit ratio of 90%, 9 out of our Top 10 movies got an Oscar. The whole study was only based on social sentiment of users in a Internet message board. We published an academic paper on this at a popular conference in 2008 (called ECIS).
This was basically the starting point for us to dig further because we saw what was possible analyzing public data.
It was nice to find correlations for the Oscars. However, we wanted something which was really useful and we could earn money with. Our conviction of the topic was really large. After some thinking it was clear that we have to do this for financial markets. I wrote my Diploma thesis about the correlation of communication in Yahoo! Finance and stock prices in 2008. The results were really promising. That was another confirmation for us that we were on the right track heading towards our idea to build a company around this. After some more months of preparation and many business cases later we got a funding from the German government (EXIST) in 2011. In May 2011 we eventually founded the company StockPulse.
How difficult was the start and what challenges you had to overcome?
Ten years ago, when we had the idea of crawling public data to predict stock prices people were really sceptical. Particularly, potential customers, banks, funds etc. were telling us that we are way too early with our idea and that they do not think that this will be relevant anytime soon. The problem was the idea was mainly in our heads and we did not have much to show. It was pretty clear that we needed a kind of prototype which we can show and explain people and customers.
Luckily, we found a start up funding program (EXIST) from the German government which supports projects in very early stages. We did not get the scholarship in the first hand because the decision committee also had some doubts concering our idea. However, after some iterations and clarifications they granted the scholarship and we got the funding. During this time we built and programmed the basic fundament of the software. After seven or eight months we came on the radar of a private equity company from Switzerland and some private business angels. Eventually, we did two funding rounds with them.
Who is your target audience?
We are mainly targeting hedge funds, asset managers or banks who use our data.
What is the USP of your startup?
We provide datasets and natural language processing in multiple languages, for example in English (which many other competitors are also doing), German, and Chinese. Non-Us/Non-English datasets or analysis in this field is very rare.
Our historic database is one of the largest available. Since we collect data for more than seven years. Most of the data can not just be purchased for the past.
We developed working trading strategies based on our sentiment data which are used in real money trading setups with great success.
Can you describe a typical workday of you?
Most time of my workday I spent for technical issues. Searching new available web sources and building crawlers, administer the database, coding other programs and algorithms and everything which is related to keep the whole infrastructure up and running. Since we are also managing all servers ourself, there is always a lot to do with this.
Addtionally, we are in regular contact with our clients and also doing sales for potential new clients.
Where do you see yourself and your startup StockPulse in five years?
We want to be the first-stop data provider of Social Media Sentiment data in Europe, having the largest and highest quality data set. In five years most likely the majority of all asset managers will use social sentiment data in some way or the other. Then we will be there.
What 3 tips would you give other Start-up founders on the way?
1. Be patient
2. Stay focused
3. Think big
More information you will find here
Thank you Stefan Nann for the Interview
Statements of the author and the interviewee do not necessarily represent the editors and the publisher opinion again.