Try Quests to earn crypto1. Introduction
In this analysis, we’ll take a look at the user metrics on Osmosis, defined mainly by daily active users (DAU), as well as weekly active users (WAU) and their transaction preferences. The dashboard will be structured as follows:
- DAU analysis: based on daily activity, considering different approaches towards what “active” means on a daily basis .
- WAU analysis: same as daily but since we can do a weekly approach, depending on the number of days per week a wallet performs a transaction, we can define “active” differently.
- Transfers analysis: after seeing different behaviors, we can look at how transfers from and to osmosis look like for said users.
- Conclusions
2. Daily active users
As expressed in the introduction, there is no clear definition on how to measure daily active users, since active is not a definition per se. Thus, first things first, let’s look at the number of users which perform a transaction during more than one day. For this, I’ve defined the first two parameters on top of the dashboard:
- Start date: where you can select the starting date for comparing the whole dashboard. I’ve set this number by default to 2022-03-01, therefore the whole dashboard will be based on this number.
- Number to show: this is merely the number of days the following chart displays. Same as before, it’s been set to 30 by default.
The following chart will introduce us to some user metrics. Since March ‘22, most users have been active (in this case, defining “active” as at at least one transaction) for two days, then 1 and 3 days around the same level, and as we increase the number of days, there are less number of wallets active during those amount of days. For instance, since March ‘22 there have been 2511 wallets active during 30 different days, which is quite a nice number, compared to the 50k wallets active during only one day.
Definition of DAU
We can now create three approaches: considering daily active as “at least one transaction on a day”; daily active as “at least two transactions”, since we saw that there was a huge number of wallets with 2 different days active; daily active as “more than 2 transactions.”
3. Weekly Active Users (WAU)
Now that we’ve seen how users behave on a daily basis, we can look at the same on a weekly basis, where we can consider active users as users which perform at least one transaction during more than 4 different days a week. This corresponds to the third parameter on top of the dashboard (num_days_of_week_active) which I’ve set as default number 4. If one wishes, they can change this number and the following section will refresh based on the chosen parameter, but the descriptions will be based on the definition of WAU as a user which has performed at least one transaction during more than 4 different days on the same week.
We do indeed see a similar pattern as for DAU; a steady decline from March to July, and from that point on a small increase. Here we’re considering users which may not be considered as active on the previous section, and I’ve plotted on the right hand side the WAU by the number of days they were active, both overall and normalized. Interestingly, if we consider wallets with transactions on more than 4 days in the same week, on average there seems to be more users transacting 7 days a week than 6 or 5 days a week.
4. Transfers
On this last section, we can also extract some insights on transaction preferences. The total number of transfer transactions (not amount) TO Osmosis and FROM Osmosis is shown on the bottom-left side. Most of them come from Atom, followed closely by Evmos and Juno. On the other side, there are more transfers OUT to Atom than to Evmos or Juno, in absolute terms.
Finally, the pie charts on the bottom-right side show, for the selected parameter num_days_of_week_active (4 as default), where do most transfers come from and go to, respectively. And in this case, 18.9% come from Evmos, 15.9% Cosmos (Atom) and 15.8% Juno. Similarly, 23% leave to Cosmos (Atom), 14.5% Juno, 7.97% Stars and only 5.19% Evmos, compared to almost 19% Evmos in.
5. Conclusions
After analyzing thoroughly the activity on Osmosis related to daily and weekly active users, the key insights shown are:
- DAU was declining since March until July, and it has stayed around the same level since then.
- Same happens for WAU, defining “weekly active users” as users transacting during at least 4 days on the same week.
- Showing a 3D scatter plot, we're able to see outliers on the number of users providing liquidity, swapping, transferring, staking and voting.
- Link can be found here.
- There seems to be more users transacting 7 days a week than 6 or 5 days a week. As we decrease the number of days a week a user transacts, the number of swaps decrease and the number of staking, voting and transfering increase.
- For the WAU criteria defined, the preferred projects transfering into Osmosis are Evmos, Atom and Juno.
- Similarly, for the WAU criteria defined, the preferred projects transfering out of Osmosis are Atom, Juno and Stars.
Following this approach, we’ve seen that for all three approaches, the evolution seems to be the same one, but with different number of wallets as we should expect. Since March, there has been a steady decline in the DAU on Osmosis, related with the whole crypto market struggling to keep users. We can appreciate a dip in June due to the halting that took place during those days, but overall the pattern for all three approaches is the same. On the other side, the number of DAU has increased since mid July, after declining a bit more throughout September.
The next chart shows some more insights on the different user profiles throughout 2022. Around 40-50% of the DAU perform between 2 and 5 transactions; 30-40% of the DAU perform just 1 transaction; 7-10% between 5 and 10 transactions, and the rest between 10 and 25 or more than 25 transactions.
I’ve managed to create a nice 3D scatter plot to try and detect clustering and/or outliers! Source code and animated link is provided on the conclusions
What we’re seeing here is a 3D scatter plot, where each dot represents a day (since March ‘22), the amount of unique users swapping, amount of unique users transferring, amount of unique users staking (or unstaking/restaking), amount of unique users voting and amount of unique users providing/withdrawing liquidity.
- The pattern we can see on the first plot on the left is directly related to what we just saw; the cluster higher on transfers, LP users and swaps corresponds to March, where there was more activity.
- On the second plot in the middle, we also see a relation with the number of swaps and LP users axis, but it is not that much related to staking users, meaning that those users staking/unstaking/claiming/restaking stay the longer! And curiously, the spot on the top side seems to correspond to an outlier on April, where there was a huge spike on user activity staking.
- Finally, the chart on the right seems to be in a middle point between the other two ones; the users voting are more distributed during time.
The next pie charts show the preference in terms of type of transactions of the users stated above; users active 7 days a week prefer swaps (50%), transfers (18%) and staking (12%) overall. As we decrease the number of days active, the swap preference drops up to a low of 34%, while staking, voting and transfer increase.