Our mission at Intrinio is to power a generation of applications that will fundamentally change the way our broken financial system works. Intrinio data feeds form the basis of everything from large enterprise business reporting applications to startup fintech apps. It’s rewarding to see our product come to life at the hands of today’s most innovative developers building powerful things.
We’re lucky to be in a business where we grow together with our customers, and we’re proud to show off their hard work. Each blog in this series will highlight a customer that has leveraged our financial data feeds to build something incredible.
Meet TickerRank, a stock recommendation engine powered by advanced machine learning. We spoke to Suraj Waghulde, founder and CEO of TickerRank, about this new approach and what inspired him to start the company.
Give us the elevator pitch for your company.
TickerRank is built on advanced machine learning technologies to rank and recommend stocks by the leaders in AI/ML technology. TickerRank ranks all stocks worldwide and provides you with stock investment recommendations based on profound scoring mechanisms for value, growth, risk, market sector, and competition.
TickerRank performs more than 100 billion computations daily to score and rank companies worldwide. You will get all these scores to figure out value vs growth stocks or best stocks in a particular market sector.
TickerRank has comprehensive machine learning algorithms to learn balance sheet, income statement, and cash flow values, which builds nonlinear correlation with stock prices. It uses Dynamic Time Warping and other nonlinear similarity measures (including SVM, multiple, and polynomial regression) to evaluate each stock and predict if a stock is going to grow and how much.
TickerRank is the first of a kind to apply the above technology to the stock market. TickerRank has a vision of providing common people with our wisdom (machine learning algorithms) of stock price prediction and helping them invest and grow money.
What is your professional background?
I was born in Mumbai and since childhood, I’ve been interested in numerical methods. I did internships at IBM Watson Research (Finance) and Google. Later, I worked in machine learning and distributed systems at Apple (Search), Yahoo! Labs (Search), Salesforce Platform, and Samsung Memory Lab for 10+ years.
I decided to apply distributed systems (big data) and machine learning techniques to investing a few years ago. Using the power of numerical methods and distributed systems along with machine learning algorithms helped me to build a stock recommendation engine.
What inspired you to start your company?
Good investment recommendations are hard to find and very expensive for retail investors. Retail investors suffer a lot due to the quality of investment advice available at cheaper prices. I followed investment recommendations from different sources and personally lost a lot of money during the financial meltdown of 2008–09 and later years.
I accumulated a unique skillset of distributed systems with machine learning, especially in ranking and graph matching field (I studied various ranking algorithms in web search, social, and temporal graph matching), along with exposure to finance at IBM Watson Research. Given that finance was my first exposure on the job, I never stopped learning about finance, accumulating vast knowledge about analyzing income statements, balance sheets, and cash flows. With this unique skillset and exposure, I decided to apply ranking to the stock market the way search engines like Google and Yahoo do for web search.
What have been your biggest challenges, and what did you learn from them?
Gathering raw data has been the biggest challenge that I faced, and thanks to Intrinio that challenge is solved. Before talking to Intrinio I did talk to Reuters and Bloomberg and their prices drove me crazy even to start building something.
Another big challenge was data quality and I give another credit to Intrinio. Their data quality is very good. Initially, I presented data issues to Intrinio and they quickly resolved them. When you’ve just started a company, how do you get prospective customers to trust you with their own money as investing comes at a large risk? This is another challenge we are facing now and Intrinio is coming to our rescue again.
What have been your biggest successes?
For our early customers, they have observed large gains (5X on few stock recommendations within a year and 2X on several recommendations, with 79% of our recommendations achieving profits since the beginning of the service) and reliability of our stock recommendations.
They see our use of techniques like SELL PUT to reduce the cost basis at volatile times. They have also appreciated and tipped us for market insights on macroeconomics like inflation, rates, and stimulus effects.
We are proud of our advisors and they are making a great difference to our product and its machine learning capability. We collaborate with a Wharton Finance Ph.D. and ex-professor for his advice. We collaborate with a successful Stanford machine learning graduate for his advice. We are also getting advice from a couple of professors at Wisconsin Madison and the University of Texas who have multiple successful startups to their name and have sold startups to internet giants.
How do you set yourself apart from your competitors?
1. Our advisors — we get guidance from successful industry and academic leaders.
2. Machine learning algorithms — machine learning is still growing rapidly in many areas. One area that is largely unexplored is finance. And in finance, the features are numerical and graph-oriented. While most (70%) currently successful areas of machine learning use gradient boosted trees and regression, others use SVM and then deep learning.
Deep learning needs a huge sample of correctly labeled data for prediction accuracy which is just not possible in finance as the correct label notion does not exist in finance. So, finance remains unexplored for machine learning and a new set of machine learning algorithms should be used.
This is precisely what we are doing — we did our research of finding correct machine learning algorithms to be used for finance and we have been combining them to enhance our predictions. Next, we went through features to be used for every stock and we currently have 1,000 features per stock (most of them are derived from financial statements), resulting in more than 100,000 computations per stock to compute its score and price predictions over time.
Why did you choose to work with Intrinio?
We chose Intrinio for their data quality and price. Intrinio filters out incorrect data which gives us the advantage as with incorrect data machine learning learns incorrectly and affects our ranking big time and we have seen that from other data providers. We cherish our relationship with Intrinio!
How would you describe your experience working with Intrinio?
Intrinio is very responsive, I never found their service was down. The performance of the service is also very good. Intrinio is prompt in resolving technical and non-technical issues.
How can people try the TickerRank platform?
Intrinio has been a large part of our success and we want to give special thanks to the Intrinio team and all its users with promotions to use our platform:
IntrinioQ — one time 80% off $65 quarterly subscription plan ($13 for first quarter)
IntrinioY — one time 20% off $139 yearly subscription plan ($111 for first year)
Learn more about TickerRank.
Need financial data to power your platform? Learn more about Intrinio at intrinio.com.
Originally published at https://intrinio.com.