Introduction

A popular way to invest in an asset over the long term is to dollar cost average. This means that investors buy an asset at specific time intervals, regardless of price, which has a smoothing effect on the average entry price and attempts to avoid short-term volatility. To answer this question one has to look at the "seasonality" of the asset, borrowing a better known term from commodity trading.

In this case, we will conduct a seasonal analysis on Sushiswap activity to see if there is a particular month where there is more activity. The analysis have been divided into two main parts:

• Global monthly activity over time
• Seasonal weekly day activity performed for the past 3 months.

Methodology

First of all, to carry out the first part of the analysis, I have taken into account the transactions included in Ethereum DEX swaps table referring to Sushiswap platform. From that, I extracted the amount in USD, the distinct from_address abd the distinct tx_id to calculate the total monthly volume, the total monthly users, as well as the total monthly transactions.

On the other hand, for the second task, I used the same table, but calculating the metrics taking into the day of the week and then, grouping by each day. This is useful to see if there are some days with more activity on Sushiswap or others with low activity. In that case, the average of the volume, number of users and transactions were computed.

Results: Global Analysis

In the monthly analysis, we can see how there are 2 months when the activity jump, and two others when the activity seems lower than others. The good months for Sushiswap in terms of volume are the May 2021 and October 2021, when the volume was around 25B USD. However, the averag swap volume were higher in May. In the case of July 2021 and Feburary 2022 (not finished yet), the volume were lower. The February is not finished yet, but we are at its final days and the volume is currently too low. In July 2021, less than 10B were moved, with an average swap lower than 11k USD.

The volume could be lower for these months because two main things. On the one hand, the cyclic trajectory of crypto market. On the other hand, because of July is a summer month and Feburary a transition month when the major of the times the price of BTC and the whole cryptomarket goes down. For the top months, the scenario is on the other way around. May is the month of prosperity and previous to the summer, when the people leads or their funds prepared for the summer. Also, in October, started the best quarter period for cryptomarket and Bitcoin, where the most percentages of times the market looks green.

The users joining pools are similar over time. However, the transactions vary a little bit more. However, we can see how the poorer months in this aspect are May 2021 and February 2022. In this case, we cannot see any pattern.

If we take a look at the final chart regarding all-in-one metrics seasonal analysis, we can see how the best month has been the October. In average, it was the month with best performances in all metrics. On the other hand, as stated before, the July and February month are the months with poorer metrics.

However, all of this results are not reflecting any seasonal trends. To do that, we need to go deeper and observe if there are some pattern in a specific period of time. Because of it is not possible to do by month, due to low minimal data, I have carry out the following task by weekly days, taking into account the average values of each metrics for each specific weekly day.

Results: Weekly Day Activity

Then, the results show the average number of active users, the average transactions, as well as the average volume by weekly day. Then, we will see which day of the week experiments the most important activity and which of them the worst.

From the three charts above, we can extract several things:

• The most active weekly days by number of active users are Monday and Friday with more than 14.5% of the users being active each of these days, the first and last day of the working week. On the other side we have Sunday, when 13.8% of the users are active.
• The average volume is higher in Monday and Friday as well with over 525M USD, but Wedneday have a good average too. The worst day is Sunday again with around 345M volume, 34% lower than the top days.
• The transactions averages are similar for all of the weekly days. However, while in all weekly days, the average number of transactions are above 12k, in Sunday, the number is below 12k. Again, the Sunday has the worst number.
• Finally, regarding the average swap size metric, the better numbers are registered for Monday and Friday with 18.5k USD and 16.7k USD respectively, but the best average is for Saturday with almost 20k USD average swap size. Again and again and again, the poorer day is Sunday, with an average swap size of 12.4k USD.

Conclusion

To sum up, we have seen how to carry out a seasonal analysis. On the one hand, we have performed a global analysis about the Sushiswap activity by month, concluding that there are month with more activity than others, but not being able to see any dynamics due to the lack of data to be able to compare an average of each month. More years of study would be needed.

Nevertheless, we have carried out a second analysis regarding the seasonal analysis on Sushiswap by days of the week. In this case, we have been able to extract information for several months and see which days are more likely to have more activity and which are less likely to have less. The analysis concluded that Monday and Friday dominate practically all facets, in terms of volume, number of users, etc. Sunday, on the other hand, is the day when there is the least activity in Sushiswap. This is normal because people tend to be less aware of cryptocurrencies during the weekend. So, these numbers seem to make a lot of sense as one tends to invest more during the week, especially at the beginning or right at the end before the weekend, to leave everything in place, rather than during the weekend, when people don't get into a routine and tend to do more socialising.