CityFALCON

CityFALCON is a 21st-century financial news aggregator. Its score rates financial tweets, news, and authors by using Natural Language Processing.

CityFALCON’s customers are investors and traders, both professional and amateurs. They also provide services for corporate clients, who want to integrate CityFALCON’s API into their own trading platforms.

Location: UK
Industry: FinTech

Product: News aggregator
The scope of our work: backend/frontend
Solutions: ML, Natural Language Processing and E-learning
Web site: https://www.cityfalcon.ai/

“Sloboda Studio is a consulting company that has helped CityFALCON.com to scale up the business very rapidly with its Ruby on Rails development services.
They delivered a high-quality product, paying particular attention to the testing phase.”

Ruzbeh Bacha Founder and CEO, of CityFALCON, UK

Client Goal:

The Client met Sloboda Studio at the very early stages of the project, and we’re happy to say we’re still working on CityFALCON’s development and growth. The story started when Ruzbeh Bacha, а former Skype employee, decided to create a financial news website. Then our future client studied Ruby and developed the CityFALCON’s MVP himself. But in time he decided that the project requires more scalability and started looking for new developers. So then we did find each other.

Our client wanted to launch a new and improved MVP with a focus on clean and simple UX to demonstrate the huge potential of his Social Media Aggregator.

It aimed to democratize the financial news industry and “Bring Bloomberg to the consumer”, giving all investors and traders equal access to financial information. As an innovative product, it posed many challenges that required flexibility and super efficient solutions.

What We Did

Client Goal:

As we started the project with already built MVP, we knew we would face scaling issues. As soon as first users appeared, the number of records in the DB increased a hundredfold, and many shortcomings of the old front-end became obvious, combined with back-end rendering and a lot of asynchronous JS.
To eliminate the scaling problems, we added several servers to the DB, reducing the burden on the main server.
We also reduced the time for requests processing and the number of requests added caching of the most frequently used data and used a load balancer for the web-servers. We took measures to filter spam bots and block useless crawlers.
Another solution that helped us avoid scaling issues was splitting the application into separate components (Enterprise API, web, processing engine).

Products Page: Consume СityFalcon news via text and voice

2. Optimizing requests processing:

Up to 10 watchlists with up to 100 financial assets. Your watchlists are synchronized between the app and web-version.
14000+ financial assets – stocks, commodities, forex, indices to choose from, and the number is growing each day. Customize the news feed as you like.
Top stories from the last day, latest streaming news, exclude/include tweets.
30+ languages.

3. API Integration:

We placed the data in the Cassandra cluster and formed the Elastic Search index for quick data output for the Enterprise API and for free of charge users. That allowed to achieve response requests for articles on more than 10 million records within ~ 100 milliseconds.

Challenges and Solutions

Challenge 1:

Adaptability to different business models.
We had to adapt to the new business models, as far as the client changed them several times.

Solution:

Agile practices.
We used stand ups, retrospective, backlogs, etc. to ensure timely and efficient implementation of all of the project’s features.

Onboarding page: Ability to choose segments to get personalized financial news

Watchlist page: List of financial news based on the client’s preferences. Ability to create numerous watchlists

Challenge 2:

Real-time big text data processing.
We had to process a large amount of information in real-time and without any system breaks. One of the overload examples was Brexit news publications. As a financial aggregator, CityFALCON did track all the news on this topic, though working at the peak of its capacity.

Solution:

To handle any overload, our developers maintained CityFALCON’s functionality in real-time. The earliest version had problems with processing of a large amount of data, but now CityFALCON can process up to 2.5 million articles a day.