We have developed a tool to watch the possibility of the market crisis for traditional finance. It is supposed to trigger an action to minimize expected loss from the situation after detecting anomalies in the market. We need to develop such a thing for the crypto world to secure the assets. Since the crypto world and traditional finance world have different characteristics, the existing tool did not detect the most recent market crash. So we have started to investigate possible data sources for that purpose. Few data analysis taught us that particular type of data seems to be valuable for the objective. We created a prototype to check the feasibility of the data. When we tested the tool, it detects the market crash successfully. We think we find a good data source for risk management. We will continue to improve the tool.
The below is a screenshot of the prototype. Green dot means high possibility of market fall, whereas black dot is for market rising.
As we are preparing MoA v2 site, the number of data for crawling has been increased. And furthermore, the performance of data crawling is getting to be a severe problem. MoA v2 site has statistics information that requiring heavy data calculation such as Korean premium, category indexes. To speed up the processing time, we applied a parallel data crawling handler. Before that, we have one big crawling handler, which manages all the crawling tasks. We run three crawling handlers for crypto, KStock, and batch processing. We will continue to improve data crawling stuffs for Moa v2.
I had a business meeting with the business partner last Thursday to discuss the implementation stuff. The partner said that they need to implement 30-40 pages to support EDN. The number of pages was beyond my expectation, but after hearing his plan, I could follow the partner.
We open a public github, https://github.com/Edenpartners2018/moaFrontend
The repository is for Moa v2 frontend. We will open more repositories gradually.