Getting the data

I first referred to terra.msgs with:

WHERE block_timestamp >= '2021-11-01' -- Kujira inception
  and TX_STATUS='SUCCEEDED'
  and msg_value:contract ='terra1e25zllgag7j9xsun3me4stnye2pcg66234je3u'
  OR msg_value:contract ='terra1w9ky73v4g7v98zzdqpqgf3kjmusnx4d4mvnac6'
AND INITIAL_ACTION = 'submit_bid'

in order to get count_unique_submits and cummulative_sum_unique_submits.

Note: These are not activated bid submissions, but just bid submissions. I was not able to find a quick way of differentiating the two.

Then I referred to anchor.liquidations to get the different liquidations and then joined the two tables.

What can be seen here is the different amounts of liquidations, both in their denom and usd as well as the counts.

Determine the total unique users placing bids on Kujira since its inception

This will be our COUNT_UNIQUE_SUBMITS and CUMULATIVE_SUM_UNIQUE_SUBMITS.

What we can see here is that since the Kujira inception date (1-NOV-2021 sourced by sam#0575), is that there's a massive spike in the count of unique submissions for participating in liquidations. A spike of ~ 10k within the first month of launching.

The current unique submission count is at ~50,272.

Compare bids on bLUNA vs bETH

Let's look at the total liquidation count from the terra.anchor table first

The following diagram shows us liquidations taking place since the inception of Kujira in both bLuna and bETH bAssets. This does not tell any story with regards to active_bids/liquidations. My attempt to getting this is shown bellow:

What we can see from the graph above is that this ratio goes above 1, which should not happen. This happens both when I take COUNT(DISTINCT SENDER) and COUNT(SENDER) divided by their equivalents for the liquidations table.

Also, what can be seen in the SQL query that yielded the above graph, is that I had tried to bring in the Anchor liquidations table to join in order to get the actual liquidation count. No matter which way I tried to join INNER, LEFT or OUTER, I was not able to get the data to match.

I am a bit lost in finding out the active bids compared to just submit_bids.

For this reason, the rest of this analysis and discussion will be focused on the first 2 graphs presented above and not this last one.

Summary

In summary, what we can see is that ever since the creation of Kujira, the total participation in liquidations in the network has increased massively. But alongside that and towards the later months, there have been other smart contract platforms such as Terra Toolbox and Lighthouse Defi, which have removed the need for the participant to pay the 0.5% in KUJI or the 1% in the won liquidated asset when they try to claim their funds. I am bringing this up as they all expose the underlying Anchor contract with their own UI, so not all the liquidations we see are directly from KUJIRA

What we can also speculate about is that the drop in participation for liquidations around the months of March until very recently, is due to the fact that LUNA was pumping, therefore, there were a lot fewer people willing to liquidate others for a 5% premium on a 110 UST worth LUNA.

Evaluation

If I had more time to devote to trying to solve this problem and this was an intermediate task, I would try to get the CURRENCY or COLLATERAL from the terra.transfer_events table and see how I could show a graph of COUNT(DISTINCT SENDER) based on what asset they used UST or aUST. Similary, I would have tried a different way to bring the liquidations per collateral with the unique active bids together, such that I can show a graph of active bids that liquidated someone and which collateral in % and in ratios.