What is Alternative Data? | The Big Explainer | Refinitiv - YouTube

Welcome to ‘The Big Conversation Explainer'.

Alternative data, alt data research and analysis, has been growing in popularity as the variety

and sources of data have increased and the computing power to easily analyze the data

has rapidly improved.

But what exactly is alt data?

Alternative data refers to data that is derived from external sources.

Essentially, anything outside of what a company typically releases.

Examples would be app usage, email receipt, employment data, geo location, satellite imagery,

sentiment, weather data.

The list goes on and on.

There's a lot of data out there.

Much of the new alternative data is based on its relevance to the moment.

So, for example, if we're looking at cell phone location data, we can see where people

are moving right now, which stores are going into.

Or if we look at satellite data that tracks parking lot capacity around the retail holiday

season, and we see that a parking lot is especially full, we know that that retailer probably

is having greater sales.

There are a couple different types of satellite imagery, one is optical, which is you get

to see objects on the ground, for example, cars or ships in the port, and use that to

gain insights.

And then there's also infrared data which gives you the ability to see lights as well

as pollution and particulate matter in the air.

Alternative data is extremely important in times like today with Covid-19, when there

may be a lack of information coming from countries like China or other areas.

Data sets in the satellite imagery space can give valuable insights around manufacturing

levels or even when looking at oil storage levels which is a topical concern these days.

So my company, MarketPsych, creates an alternative dataset that's based on the analysis of the

information flow, where we digitize how people speak about assets like currencies or commodities

like crude oil or gold, or 15,000 stocks or stock indexes or even sovereign bonds.

We're able to pull out all of those references and create sentiment time series, as well

as emotional time series and macro economic time series about each one of those.

And it turns out that if you, for example, aggregate their overall sentiment, their overall

positive versus negative feeling about a given stock, that does have some predictive power

into the future.

You know this is a lot of real time information

that's coming through that investors can use.

So, again, you know, from a data standpoint it's very exciting because of all the different

types of data coming through.

And the real challenge there is exactly how you use it in your research or investing workflow.

You know, you got to really understand very well, you know, what kind of value that data

is providing to you and how it fits into, say, your economic model for how, you know,

how value is created in the economy.

And with an understanding there, you can make, you know, investment decisions.

I think it's really about covering blindspots.

So one of the things that I find most useful in the traditional data world, is looking

at what the sell side analysts do.

They're the professionals that know a handful of companies really well, and they have a

pretty good sense for what those companies earnings and other financials will be at the

end of the quarter.

For me, coupling the sell side analysts views with alternative data is really the sweet

spot.

Corporations use alternative data as well, but sometimes in different ways.

They may use it to check their supply chain, make sure there's no disruptions, or they

may even use it to try to understand brand perceptions.

By observing the light, we can get an indication as to the level of activity and how much work

is being done and how much product is being produced from that individual facility.

Now, most alternative data is also not only real-time, but behavioral data.

It's about what are people doing?

What are they spending their money on, or what are they concerned about?

So when I talk about concerns, I'm really going into this area of information analytics.

Like what are people worried about in social media?

Are they worried about a new epidemic virus or are they worried about interest rates and

deflation or inflation?

These are things that we can pull out of media, social media or news media, and we can digitize

using natural language processing.

I did some analysis with LinkUp, that's a alternative data company that scrapes job

postings from the company's own websites and tracks say active jobs over time.

And I looked at that alongside I/B/E/S, which is a Refinitiv product that captures those

analysts views.

And I saw several situations where companies were pulling back job postings probably a

few weeks, a week or two before the analysts revised their earnings estimates down.

And I think that's really part of the key with alternative data.

We've seen the most valuable signals occur when different clients tend to start combining

data sets to get a stronger signal.

You really need a very robust research goal in mind when you engage alternative data.

Because there's so much of it out there, you need to have a really good sense as to which

data sets you want to engage, you know, how much you're going to rely on them in your

investment workflows, and then ultimately how you know how that's going to impact your

portfolios, your investment decisions in the market.

We will be delivering our first alternative data set through QA Direct in the cloud and

that data set will be LinkUp, which is the job listings provider, and LinkUp data, it's

been extremely useful as of late in helping the users of LinkUp data to determine which

companies or which industries have been hardest hit throughout this this Covid pandemic and

as well, you know, which industries from a from a job listings perspective are starting

to kind of bottom out or starting to recover.

To summarize our involvement with BattleFin and the alternative data space is largely

based on Refinitiv providing content to that community such that the community can really

easily prove up the value of that alternative data, and as well delivering portions and,

you know, subsets of that alternative data downstream to our customers.

Alt data's clearly been useful in analyzing sentiment and behavioral trends around the

spread of Covid-19.

But it has also stepped into this breach during the period in which traditional corporate

and macro data has been incomplete or distorted by complications around methodology during

periods of lockdown.

Analysis of credit card data, hospitality booking systems and satellite imagery may

even supersede some of the existing forms of analysis because the alt data can be gathered

in real time and can't be hidden in the footnotes of corporate accounts.

For investors searching for real time edges, this type of data will become an increasingly

attractive alternative.