Frequently Asked Questions

What does your data analytics include?

Our main proprietary data analytics combine attention volume, attention buzz, sentiment, events and influencer information. Users can drill down e.g. to identify the sentiment for a certain event or attention from a specific influencer group. Information is also aggregated to the level of industries, sectors, indices, countries and regions.

How do you deal with fake news?

We only focus on relevant content from relevant sources. Sources are qualified, classified and scored to identify the relevant ones. Also the content goes through a qualification process, during which the proprietary content relevance score (CRS) identifies the financial relevance of news articles, blogs and tweets based on criteria such as novelty, reach and originality. Read more about this process in our Sentifi Engine whitepaper.

How do you deliver the data?

Our data analytics are delivered via API, desktop (web platform) and customized widgets.

Does your data generate alpha?

In-house and third party backtests prove that Sentifi’s data can be used to create alpha-driving trading strategies that outperform major indices. Check out our 2-pager for quantitative asset managers.

Why do you use influencer scores?

Not every source has an influence on the valuation of an asset. We therefore measure the influence of sources with our proprietary influencer score.

What’s your current data history?

Currently, we have 1.5 years of history. The data history will be extended to 3 years by the end of November 2019.

Do you offer trials?

Yes. Please contact our Sales Support and request a trial account for our intelligence products.

Where do I get more info about the price?

Please contact our Sales Support for information about product prices.

Do you use AI methodologies?

The Sentifi engine combines a multitude of AI models to structure unstructured data and generate unique analytics that enable actionable insights. Read more in our engine whitepaper.

Do you use deep learning models?

For each problem, we use the most appropriate AI model. Deep learning is a popular methodology, but not always the right choice. Currently, we use deep learning for influencer qualification, influencer classification, bot detection and sentiment detection.

Can data be shared with 3rd parties?

To share data, the 3rd party needs to sign a non-disclosure agreement. Contact our Sales Support for further information.

How many assets do you cover?

We cover approximately 50’000 assets. We track all listed stocks globally, all currencies (including major cryptocurrencies) and all exchange-traded commodities. In addition to that, we cover industries, sectors, indices, regions and countries.

Do you track cryptocurrencies?

Yes. We track all major cryptocurrencies including Bitcoin, Etherium, TRON, Ripple and over 100 more.

How is the sentiment score calculated?

The sentiment score is calculated based on all news articles, blogs and tweets mentioning an asset during a selected time frame.

How are attention volume and attention buzz calculated?

Attention volume is the number of news articles, blogs and tweets written by influencers during a selected time period. Since companies like Apple, Google or Facebook would always be on top when measuring attention, we have developed the metric “attention buzz”, which compares the attention volume for an asset with normal attention levels. This makes it possible to identify outliers i.e. companies that get unusual high attention.

How frequently is the data set updated?

Data is processed in real-time as it comes in and updated in our analytics every 15 minutes. Since we use EOD prices, price data is updated daily. Asset allocation data is updated when funds reports are published, which depends on their reporting frequency e.g. monthly or quarterly.

What are your information sources?

We don’t scrap any data. Prices, company information and industry classifications are sourced from Morningstar. Financials and asset allocation data is sourced from Factset. News articles and blogs are sourced from LexisNexis and RSS feeds. Tweets are directly sourced from Twitter. 

What differentiates Sentifi from other alternative data providers?

Our sources are qualified, classified and scored. We have qualified more than 200 Million sources to identify 14 Million financially relevant influencers, whose content impacts the valuation of assets. Also the content goes through a qualification process, during which the proprietary content relevance score (CRS) identifies and scores the financial relevance of content. Sentiment detection does not focus only on words but on the impact on an asset and is target-sensitive. The message “Apple wins lawsuit against Samsung” will assign a positive sentiment to Apple and a negative sentiment to Samsung.

How do you define the term "influencer"?

An influencer can be a person (such as a CEO) or an organization (such as Reuters). Influencers are defined as those sources, which content can affect the value of an asset. We have qualified more than 200 Million sources to identify 14 Million financially relevant influencers.