Crypto adoption, social media and influencers

Transcript

okay hello everybody well thanks thanks a lot for coming today pleasure to be at sec again this year um so today we're going to go and delve a little bit in the wonderful world of social media influencers and what's going on in the crypto market um so before we we start one thing that this talk is not going to be about is obviously investment advice there's no investment advice and i would like to acknowledge a lot of the work i'm going to show you comes from eric any mistake is mine and a lot of that also comes from our labs team and our data science team so let's get started um so

what we'll try to look at today is okay what is affecting crypto right do we have any fundamentals and if and not in the market voter fundamentals what is driving the price of crypto assets and what is driving engagement how people interact with those protocols right um so we're going to try and look at different data sets to answer to answer those questions uh trying to understand how opinions form on social media uh we'll have a little glimpse at elon and especially recently with twitter so we'll have a look at twitter data for elon and finally we'll try to answer the question

can you trade on the sentiment of the twitter chatter so we'll start with the data sets uh to set to set of the stage then we'll go into what how can you measure crypto adoption through social media there are a lot of other metrics you can look at but we'll focus on social media for this talk and then we'll delve deeper into influencers and elon is he an influencer is the influential tbd and finally we'll we'll end with the impact of news right in in the traditional markets when news get published on bloomberg thompson reuters and other

mainstream publications often you see an impact that's very quickly incorporated into prices do we have the same in crypto is there an impact over what timeline and on what type of information and finally we'll have a little bit of fun and just try to see if there are ways to trade on that information that is present on social media so data set uh so before we start just a little bit about about us and and what we do so i'm sasha from uh the thai.io uh so we're an institutional service provider we have our main product is the front end and we provide institutions with all the information

that they need to operate in crypto focusing on market moving information so hence the the idea for this talk and we'll focus on a subset of of our data sets including some market data very simple market data like prices market cap what we call significant developments so these are our proprietary tags that you can see on the right so these basically show you go to thousands of sources in real time these sources are completely crypto native so we're not just looking at a yahoo finance or bloomberg we're also looking at medium articles and blog posts um and for each

news that come in we'll get a tag we'll get the coin that it's associated with and we'll get those and read significant developments that tell you uh kind of corporate actions for cryptocurrencies uh the bulk of the of the data set i'll i'll talk about today obviously is the twitter data set uh we do a lot on twitter uh first obviously we have the tweet volumes that are that incorporate a pruning of bots so we try to filter the noise uh by by removing a lot of the bots um maybe elon wants to know a bit more about that we also do a sentiments course or in each tweet we'll go and rank each word to understand is the sentiment positive

or negative we have a bunch of models for this and number of followers geolocation of the accounts that posted the tweet and finally a bunch of other social media sources just that to make things a bit more interesting including including youtube so what is crypto adoption before we get into um influencers and and get closer and closer to something that looks more like fundamentals step by step uh what is crypto adoption one thing you would think about crypto adoption is how many people hold cryptocurrencies uh well in that room i'm not gonna ask the question um but you could also look

at uh how many people are interested in looking at cryptocurrencies or you could look at how many people talk about cryptocurrencies so here is is a snapshot from one of the tools we've built to look at the distribution of tweets across the world by different tokens so here we're looking at the distribution of um twitter activity on ethereum across the planet over the past month and if we zoom here on the on the little chart on the right you will see that in france something is happening this week the interesting thing if you zoom in

france is you'll see that the top coins are actually gmt so i guess people are doing a lot of exercise in paris btcb which is interesting in light of all the announcement that have been going on recently with with binance and the amf uh tezo's uh french shout out and uh and obviously celsius and the current market but we won't say anything about that now this is on a global scale so to try and focus we'll go a little bit further not too far outside europe so let's go to the uk and see how crypto adoption looks like in the uk so one thing that's important is to obviously normalize the number of tweets

that you see by the population of each region so if you look within the uk over 2021 you see a lot of activity in southern england not as much per capita further north and if you look at ethereum you will see that obviously everything is dominated by london there is somewhat of a not close correlation between london and other uk regions but you can see that there are a few peaks obviously that move the market that you would also see if you looked at bitcoin but let's look at a meme coin just for fun so how about dutch coin same you can see

that london is crushes all the rest of the other regions one big difference though is that if you look at other uk regions for ethereum there's still activity on in all the other regions across time when you look at dilt coin it's completely correlated around those big peaks anyone care to guess what this big peak is you're right elon musk on snl so as you see the diff the the the behavior on social media for meme coins and for more quote unquote fundamental coins or coins that have more utility

is quite different and if you look at a moving average of those for bitcoin ethereum and dogecoin what you'll see is again when there are phases of high activity everything starts activating itself on dutch coin whereas it was already activated on ethereum and bitcoin what's interesting here is to notice that all the the peaks the large surges in activity often started on dosh coin in 2021 so then you first had the bumps uh for dogecoin and then it propagated through ethereum and and and bitcoin and it's not causal but it's interesting to observe and if you look at those different coins across the different geographies what you'll notice is

bitcoin and ethereum pretty similar and dogecoin is quite different in a sense it shows that you know dogecoin responds to different types of stimulus than than other cryptocurrencies in being a meme coin so speaking of dogecoin first step in the market that has no fundamental is you look at okay what are people talking about and how are they getting influenced so that's that's why we can look at influencers uh one great place where people get information if you listen a little bit to all the discord channels in crypto is youtube uh so you have all those youtubers that publish a

lot of content to explain democratize and make all the scripture projects more understandable so as expected if you look at the number of average views versus market cap there is a clear correlation the bigger the market cap the bigger the interest the bigger the number of youtube views that those top videos get but how does the evolution of an influences following evolve over time what's very interesting there so here we're taking a couple of random influencers and we're looking at their following over time what you'll

notice here are a couple of things first if you look if you look at the top blue line you see that it's completely outperforming its peers uh and by even the rate of increase of their following is much higher than the rest you see a sort of mix of three influencers in the middle and then one one that's struggling to keep up what's interesting there is those jumps every time you get big jumps and that's the way influencers gain traction you get a viral post you get something that is of a lot of interest and suddenly all of a sudden your number of followers

increases and it seems to be quite sticky so it's not linear it's not about gaining followers over time it's about getting those pieces that are going to boost the number of people who are going to start following you so on to the awaited question of who is the greatest influencer or worst who knows elon of course so just a first quick look at what happens when elon tweets if you look at bitcoin well here is a snapshot of uh of the the past year he hasn't been that active on on bitcoin but you can see around uh beginning of 2021

uh was it one day i think they do you know something around tesla buying bitcoin or offering bitcoin payments uh well but more or less what you can see is there's there's a lot of tweets hard to see much on on that chart so let's move over to dos coin that's a kind of a different picture right uh so he's been quite active uh for for that period and then in after january 2021 april may as as we pointed out a few slides ago and then it exploded still not the most convincing um way to see impact on on prices so what

about if we looked at this and i'm not going to give you the answer or say that for sure elon impacts prices the only few comments i will make on that chart is what this shows basically by windows of 12 hours it shows you the maximum return so the maximum relative change in price of dogecoin so basically a big peak means there was a big spike up of the price a negative peak means that it was only a down market uh and the time at which people lost the least amount of money over the 12-hour period was around -10 what you'll see here is the magnitude of

the the change the top changes over 12 hours was much higher uh over 2021 that it is now and there was a lot of tweet uh on that on those facts around then moving to more fundamental aspects news new information should move markets as we expect from traditional finance is that the case in crypto so we delve into the um the significant development that we gather and tag uh automatically on all the thousands of sources and basically what we're going to do is look uh at the abnormal return so basically on average every time a piece of news is published what is the reaction on price 24 hours later

and if you look at this you basically center around each event and look at the evolution of price of the tokens that are mentioned on this piece of news and see if there is a consistent pattern in in the movement of price so what you see here is basically the most negative on the left and the most positive average movements over 24 hours and you see error bars are important to have an idea of the distribution of those uh changes in prices to to really know and these error bars by the way are not on the distribution of the movements they're on the distribution of the average uh return so what you can see here is what moves in a positive way is mainnet updates

network status updates listings and many activity obviously it's not clear for add drops which is interesting there may be some additional filtering to do and negative any type of financial earnings stablecoin dpegs obviously having is initially one delving into it further so can you make money trading on sentiment one thing you can do is let's get the top market the the top 100 um coins by sentiment the one that have the bigger the bigger sentiment the highest core this is what you'd see so recently you

see it's not it's not doing really well especially in 2022 another way you could look at this is and by the way this shows the cumulative return so i'm starting with a portfolio of one dollar and then over time you see how much my portfolio evolves so here you get a top of 6x on your portfolio and basically you're more or less a little bit more than even if you had started back uh early 2021. um what if you took only the top 20 or on either 21st to the 40th etc right so these were we're looking at quintiles of of sentiment over time and similarly you know the top one

performs really well but it would be you would be really really down if you had started at the end of 2021. how can we make this better and that's where the beauty of being able to go short uh or to to short an asset is is interesting so what if now we took the top performers in terms of sentiment and actually shorted on the other leg the we would fund that strategy by shorting the bottom performers and in this case what happens is you get this can we improve this well maybe now instead of taking the top minus the bottom i'm you know i i want

to look at still the top but i don't want to be completely dependent to the bottom because i don't know some of them are smaller some of them might actually get spikes in prices that i can't control that are not related to anything truthful and sentiment so what if i just short the basket of top 100 by sentiment and finally what if now i did a mix of actually getting the top minus the bottom but in some cases if i'm not happy with the top 20 and i don't think they have enough potential i'm going to take the whole top 100 by sentiment and still show the bottom because they're the ones

who are underperformed now how do we compare those strategies in traditional finance you look at risk adjusted returns you're going to look at what's called a sharp ratio typically your sharpest above one is good below is not that great the snp for quite some time was around 0.7 i don't know these days um and uh so here what as we can see the best strategy is this number three where you do a mixed approach of what type of portfolio you do with rebalancing obviously this doesn't take into account any fees any rebalancing cause any market impact or any cost of funding your positions uh for the especially for the short leg or nor about the ability of being able

to play that short lag but i thought it was interesting to see that there may be things that you can do with sentiment data i'll open it for questions and phil always i love talking about those types of social media or broader data questions so feel free to reach out [Applause] any question anyone want to share their favorite coin good thanks everyone

you