MarketPsych Posts Cryptocurrency Sentiment Indexes

The data will be available directly from MarketPsych from February 1 via APIs, and will also likely also be available via Zug, Switzerland-based data marketplace operator Streamr in March, once the vendor has integrated the data into Streamr’s platform, says MarketPsych CEO Richard Peterson.

The indexes cover the top 100 cryptocurrencies by market capitalization, and include scores for 43 different types of sentiment, ranging from general sentiment and factors such as optimism, fear, uncertainty, price direction and volatility to cryptocurrency-specific factors such as adoption and adoption forecast, criminal activity, innovation, sentiment relating to a cryptocurrency’s code or development team, transaction speed, whether a currency is likely a scam or a potential target for a crackdown by regulators, “forks” in the blockchain, and even one called “noobs,” which monitors the level of “newbie” or naïve investor activity—based on the perceptions of people talking about them online.

Peterson says potential users will include those who actively trade on cryptocurrency exchanges and want to forecast price movements and identify trends, as well as those who want to use the index factors as screening tools to identify which cryptocurrencies are likely to be a scam, or which demonstrate actual innovation, as well as institutional investors who may be unwilling to invest directly in cryptocurrencies, but want to be aware of the space as it evolves.

“We have a prospective client who already subscribes to our other datasets via Thomson Reuters and is very eager for this. And people tell me there are a lot of hedge funds in China trading cryptocurencies. They say ‘If you build this, I can sell it’,” Peterson says.

Certainly the space is evolving quickly and growing fast, with more than 1,300 cryptocurrencies already in existence—the best-known being Bitcoin—and total market capitalization of over $750 billion, and trading volumes accelerating on crypto exchanges such as Bitfinex as the market attracts both professional and retail traders.

The pace of this growth makes reliable analysis of the market even more important, Peterson says. For example, if someone hacks a wallet or exchange, investors might find out by themselves within a few minutes—but in that timeframe, the price of currencies could have moved significantly.

“We have developed analytics to look at how people are discussing these currencies in the online communities and chatrooms that cover them. We monitor all types of references, from vulnerabilities in code to hack attacks… for people who can’t monitor the hundreds of chat rooms that are talking about them,” Peterson says, noting that building the sentiment indexes for cryptocurencies was a whole new challenge compared to MarketPsych’s existing coverage of traditional financial markets and their taxonomies.

“Sentiment for cryptocurrencies is different. Obviously there are negative words, like ‘bad,’ but there are also more sophisticated terms, like ‘vulnerable’ code or ‘compromised’ code or ‘wallet,’ or phrases like ‘can’t be used as payment’ or ‘slow transaction speeds.’ And the majority of cryptocurrency users are not native English speakers, so—even though these groups have their own terminologies—the language can be somewhat distorted,” he says, adding that recognizing the most useful terms to follow is “an iterative process.”

In addition, the vendor has to deal with the “cesspool” of extraneous information on Twitter that can mask valuable data. “We have spam filters to filter some of it out, and bot detection tools to try to identify bots, spammers and marketers—for example, they tend to put in a lot of tags to get onto as many peoples’ feeds as possible, so we block those,” Peterson adds.

Though much of the interest in cryptocurrencies is recent, MarketPsych has actually had an interest in the space since 2010, and has been analyzing cryptocurrency data since 2013. “In 2010, I was in a coffee shop and saw a guy coding, so I asked him what he was working on, and it was Ripple, a digital currency. He [Diego Gutierrez] later joined MarketPsych as CTO, and so we thought we would see if we could track and predict prices,” Peterson says. “And in 2013, it seemed that there was a correlation between prices and the number of mentions, so we started analyzing and testing that.”