A detailed overview of our Products & Services

Datasets NLP Signals Strategies Fundamental & Technical FAQ

Selection of SourcesRegionMessage Type
Xueqiu.comChinaForum Posts
Taoguba.com.cnChinaForum Posts
Stockhouse.comUSA/CanadaForum Posts
Finanzen.netGermanyForum Posts

Additional Specs

Messages per day
(relevant and filtered, mapped to tickers) 
More than 2 million
(social media users, official accounts)    
More than 60 million
Total amount of messages
(relevant, mapped, and filtered)
More than 3 billion
Meta Information per day
(tagged tickers, key events, key persons, etc.)
More than 4 million
Historical data
(our Crawlers are running 24/7 since 2010)
10 years for US and Europe
5 years Asia
Amount of Data Processing per month
(our AI detects the relevant parts)
More than 15 Terabyte
(supported languages in NLP)
English, Chinese, German
Major Indices
(S&P 500, Dow Jones, Nasdaq, DAX, TecDAX, etc.)
(Gold, Silver, Oil, Copper, Palladium, etc.)
(Materials, Industrials, Financials, etc.)
(Electrical Equipment, Professional Services, Airlines, etc.)
Key Persons
(Executives, Board of Directors)
More than 250.000 names

Aggregation and Data Frequency

Defines the time unit and is chosen by default according to the customer’s selection. All available partitions can be retrieved by the API endpoint: GET /reports.

The data for the listed datasets above is processed in real time and/or near time (depending on the lag of the data collectors). There are different snapshot times where data is aggregated for different time intervals.

Default Snapshots

24 hs


10 min

Real Time


Bloomberg Ticker
Figi Composite Code
Ticker Symbol 



Buzz is a measure of volume or intensity of communication. The buzz indicates the level of attention of a title. We express the buzz value in percentage points. The buzz value refers to the number of opinions, tweets, and messages collected.
100% means average. A buzz value of 1000% shows attention 10 times above average. Our system calculates the buzz value in real-time. Each subsequently collected message or post increases the buzz value regarding the corresponding average. The point of reference is always the buzz value of the previous day and time. 


Sentiment represents the content of messages, tweets, posts, and market participants’ opinions of a stock. Our fully automized system scans Social Media for relevant opinions and messages permanently. It instantly analyzes and classifies the messages, separates factual information from spam, and evaluates content and sentiment. Stockpulse schedules the sentiment on a range of -100 (very negative) to +100 points (very positive).
All real-time data is processed and updated subsequently.


Stockpulse applies latent dirichlet allocation (LDA) to each message it collects.

LDA is a method from the field of sentiment analysis to cluster topics in arbitrary texts. 

We are also using BERT to cluster topics and identify important key words and phrases in communication.

Further links:
LDA (https://en.wikipedia.org/wiki/Latent_Dirichlet_allocation)
Bert (https://towardsdatascience.com/topic-modeling-with-bert-779f7db187e6)

Author Statistics

An author is either a single Social Media user or an entity like a company, trust, or organization

Information about authors of messages, e.g. their impact score (a proxy for reputation) or their average message counts


Persons (executive personnel of listed companies)

Products (of listed companies)

Stockpulse screens individual messages of persons and products and provides information on all entities listed.

Key Events

Key events are special events related to financial markets like central bank rate decisions or quarterly earnings reports and merger announcements. Stockpulse tracks additional key events in Social Media in more than 180 different categories.

Stockpulse’s Specialized NLP vs Google NLP API

NLP is at the core of Stockpule’s know-how, which differentiates us from our peers. Apart from collecting editorial content and key events, Social Media posts and comments are swift and short. Their length is mostly limited between 100-300 characters and represents around 90% of all collected information. To evaluate and classify these “informal” comments, NLP must be adjusted to Social Media to accurately achieve processable results.  Stockpulse developed a unique system for financial markets. Supported languages are English, German, and Chinese (traditional and simplified).

Below we illustrate the methodology of our specifically adapted NLP and ticker matching for a sample social media message:

“So my TLRY Yolo 42C 2/12 is gonna pay is what you’re telling me?”

Our NLP and sentiment analysis detects TLRY (Tilray Inc., a Canadian cannabis company) with a neutral sentiment. The term “42C 2/12” is probably referring to a call option of the user, which would indicate a positive meaning. However, the question form of the sentence leaves some uncertainty and that is the reason why Stockpulse’s NLP labels this message with a neutral score.

Google‘s NLP API detects:

Google’s NLP detects a negative sentiment of -0.4 on a scale of -1 to +1

This example demonstrates the different NLP approach between Google and Stockpulse explicitly. The adapted use case and domain have not been correctly addressed, although Google is widely seen as the industry’s benchmark.

$TLRY $APHA calls since 1/5/2021 got me to over $500k today. Still think it has much upside

Our NLP detects the ticker symbols TLRY (Tilray Inc.) and APHA (Aphria Inc., also a cannabis company). The text receives a positive sentiment. The words “calls” and “upside” are the major factors. Especially the word “call” or “calls” is a very specific term in financial markets and normally means to buy an option on a stock, which can be associated with a positive meaning.

Google‘s NLP API detects:

Our system classifies the message for both companies ( Tilray and Aphria) with a positive sentiment. (+0.3)

Google labeled it negatively (-0.3).


Signal Rules

Stockpulse offers a surveillance service that tracks buzz and sentiment for one or more titles and triggers events if a critical buzz level is reached. Customers monitor Emotional Data Intelligence in real-time. API users choose their titles and adjust event triggers by default.

Pulse Picks

Trading signals are generated with machine learning AI, referring to daily buzz and sentiment data and daily quotations. Signals are created each trading day approximately one hour before regional market open time.



This strategy has been very successful since 2011. Stockpulse customizes equity indices according to the client’s demand.
We cover European, American, and Chinese indices. Additionally, ESG, Technology, Growth, and Value indices. Stockpulse selects a benchmark equity index and rebalances the index. We offer monthly, quarterly, and semi-annually rebalances based on our retrieved signals.

Long / Short

This strategy has achieved amazing results since its implementation in 3/18 for TSLA, AAPL, and the S&P 500 Index.

Our deep learning artificial intelligence retrieves buy and sell signals from Social Media to establish successful trend-following strategies.

Market Neutral

We implement balanced market-neutral positions for the S&P 500. The strategy uses MOC (market on close) signals to open and close stocks on the long and short sides, holding them overnight till the market closes the next day. While trading volume spikes and spreads tighten during the market close, Stockpulse’s deep learning AI, based on TensorFlow, automatically detects and selects the strategy.

Fundamental & Technical Data

Listed Companies

Indicators based on price data and fundamental key figures

  • Traditional financial information such as earnings releases, analysts updates, dividends, etc
  • Titles close to provided value for their relative strength index (RSI)
  • Price based simple moving average (SMA) for titles in the given span
  • Price based simple moving average data for a single title
  • Historic short sales for a title and the chosen authority: FINRA (U.S.A.), FCA (U.K.) or Federal Gazette (Germany)

Specific Earnings Data

  • Titles which have earnings releases during the next 2 days or in any given time frame
  • Price ratios (e.g. P/E) for a company, its competitors from the same country and worldwide

Price Data

  • Stockpulse monitors trading logs of a variety of freely available sources
  • From these sources Stockpulse calculates a real-time indication as well as minutely, hourly and daily OHLC values

Short Interest Data

  • Stockpulse provides data of short positions for US, UK, and German equities

Unlisted Companies

We cover over 3M unlisted companies from the D-A-CH region (Austria, Germany, Switzerland) and the United Kingdom.

German Annual Reports (Bundesanzeiger)

  • Annual reports of all registered listed and unlisted companies in Germany
  • Complete history since 6/2017
  • Extracting values from annual reports for subsequent results, e.g. EBITDA
  • Developed ontology for financial key words in annual reports
  • Applicable filters for all reports available

Commercial Register Announcements (Handelsregisterbekanntmachungen)

  • Notifications of executive personnel, capital increases, rights issues and relocations
  • Structured network data to identify entity

Additional specifics from Stockpulse

  • Vacancies (valid for unlisted companies in the D-A-CH region)
  • Follower statistics from Social Networks, e.g. Xing (German speaking business community)
  • Data from Employee/Company rating platforms, e.g. kununu in Germany

Application Programming Interface (API)


  • REST is a web-service protocol that lends itself to rapid development by using everyday HTTP and JSON technology
  • Our API docs provide code snippets for Python, PHP, Java, and more languages for each request

The API is hosted at https://api2.stockpulse.de/v5

  • Requests are required to use TLS/SSL
  • Returns are in JSON, XML, or CSV
  • Timestamp format can be chosen globally (Unix, ISO, UTC)
  • We offer trial periods with option to download historical data for internal evaluation


Clients can use our Websocket API to receive real time updates of our various data streams. We provide wrapper classes for some common programming languages like Java and Python, to get you started quickly.

Available Streams:

  • Buzz and sentiment tick updates
  • Triggers from signals rules (buzz patterns, user messages)
  • Trading opportunities and signals from strategies


What do you exactly do?

Stockpulse is a data analytics company offering AI-driven data and signals for financial institutions.
We screen Social Media platforms worldwide and process our signals for the financial community, media outlets and executive committees.

How do you do it?

Our proprietary software is capable of converting unstructured data into comprehensible information. We are operating a crawling framework and collecting data from several thousands of sources worldwide. Our crawlers are running 24/7.

What products and services do you offer?

We offer our-state-of-the-art machine learning AI solutions for the financial industry. 
We cover the whole value-chain of emotional data intelligence for listed companies reaching from single datasets up to index strategies. Our AI is capable of providing information even for non-listed companies from the D-A-CH region.

What kind of clients to rely on Stockpulse’s services?

Banks, Asset Managers, Hedge Funds, and Insurances
Private Equity, Venture Capital and Family Offices
Publishing Houses and Media Outlets
Business Consultants and Advisory Groups
Stock Exchanges and Financial Conduct Authorities

How do you ameliorate equity indices?

We use the asset allocation of emotions to improve index performance. We analyze Social Media squawk, filter the emotions and convert them into buy signals.

Do you offer retail products directly or over B2B partners?

Stockpulse offers only B2B services. We offer retail products via business partners. Further information upon request.

How many companies do you monitor?

We monitor over 35000 publicly listed companies worldwide. Furthermore, we provide unique datasets of 2.3 million unlisted companies from the D-A-CH region.

What regions do you cover?

We cover North America, Europe, UK, Greater China, Hong Kong, Taiwan, and Singapore.

How is the product published and distributed, and how do I get this information?

Via web-based dashboard and API. The web-based dashboard provides a graphical user interface and visualizes the data from different perspectives.

How is the company’s data used and collection permissioned from the primary sources?

We collect and process publicly available data from public domains. The data we obtain does not contain personal data or data which would identify a specific person. The user data is anonymous on a pseudonym basis. We always consider official terms of service of the source. We never use scripts to provide usernames/passwords automatically to log in to web pages.

Do you offer training and support to evaluate and interpret data flow?

Yes, we offer training and support during trials and provide an ongoing support for our customers.

What makes this data unique from what is available publicly?

We provide the emotional value of modern asset allocation. The analysis of emotions in unstructured texts is unique.
With means of NLP, AI, and Deep Learning we enrich the data to make it easily consumable.

How long is your data history?

Our data history is collecting data since August 2012 for Europe and America, and for Greater China, and Singapore since 2017.

What is the data frequency?

Our data frequency is in real-time for single messages and is adjustable to the most common scales: Monthly, weekly, daily, hourly, and 10 min.

What are the data delivery time and delays?

The data is in real-time. We provide a RESTful API and a Websocket to push the data to you instantly. The data delay is within a few seconds. We also provide different snapshots of the data, on a 24 hours (daily), hourly, and 10 minutes base.

How do you identify the 35k publicly listed companies?

The main identifier is the ISIN. Alternative mapping for Bloomberg tickers, Figi, Sedol, and US ticker symbols.

Which asset classes do you cover?

We cover Equities, Equity Indices, FX, Commodities, and Cryptocurrencies.

Which languages does Stockpulse support?

Stockpulse’s AI understands German, English, and Chinese (traditional and simplified).