Here's some specific questions I've personally been interested in, but haven't yet had the opportunity to explore:

• Can we look at data from 2007-2008 to attempt to predict the magnitude and duration of the global economic impact?
  
• Can any of our data sets (GeoRev, RBICS Focus, etc) help us understand the interconnectedness of various industries and supply chains, and predict which ones will be most severely impacted?
  
• Can data from markets with more advanced infection rates help predict similar economic effects in other markets?
  
• Can we compare different governments' responses against ongoing market or other data, so that governments can make more informed decisions about how to respond?
  
• Could analyses like these help individuals make more informed decisions about their investments, rather than panic-selling or buying back in at a bad time?

It's important to avoid contributing to the ongoing crisis by making strong claims which prove to be incorrect, so any attempt to effectively address questions like these will need to both come from and communicate a healthy posture of epistemic humility (now even more so than usual!); but I am confident that thoughtful analysis of reasonable data can be tremendously beneficial.

This looks like an interesting paper: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3570280

Quant Bust 2020
Zura Kakushadze

Abstract
We explain in a nontechnical fashion why dollar-neutral quant trading strategies, such as equities Statistical Arbitrage, suffered substantial losses (drawdowns) during the COVID-19 market selloff. We discuss: (i) why these strategies work during "normal" times; (ii) the market regimes when they work best; and (iii) their limitations and the reasons for why they "break" during extreme market events. An accompanying appendix (with a link to freely accessible source code) includes backtests for various strategies, which put flesh on and illustrate the discussion in the main text.

(First submission ever :])

Hi everyone, here is a first notebook for visualizing COVID-19 effects for factor returns and mapping related areas. Everything here as to be taken for information purpose only.

Summary:
I added the cumulative changes of reported cases to Thomas Wiecki notebook sample to visualize the stagnation of the epidemy (China for instance).
The dendrograms reflect the degree of correlation between reported cases. The standard deviation of factor returns doesn't seem to be affected by the reported cases until cases in Europe and US were rising.

However, the factor returns seem to hold well even in stress period as they slowly came back to normal.
The spread of factor returns seems to indicate higher spread than normal level, further steps will requires stationarity test of integration order in larger dataset to explore whether the relationship holds in the long run.

Take care and stay at home.

Lucas BL

@Lucas BL: Thanks for your submission -- it looks great! I think studying volatility and factor spreads is a great direction to go in. I also think further study of market correlation patterns could be very interesting.

I wonder here:

delta_df = covid_df.pct_change().rolling(1).mean().mean(axis=1)  
delta_df = pd.DataFrame(delta_df).dropna()  
delta_df.columns = ['delta_df']  
delta_df['Volatility'] = delta_df.rolling(2).std().mean(axis=1)  
delta_df = delta_df.pct_change().dropna()  

Did you mean to take the pct change of the pct-change of the covid-numbers and the pct-change of the volatility of that change?

@Thomas Wiecki, thank you for your feedback and mapping this blackmagic line. It actually doesn't affect the elbow curve but the last line should be deleted for readability purpose.

To make the code more explicit:

global_total_cases_df = pd.DataFrame()  
global_total_cases_df['mean'] = (covid_df.pct_change().rolling(1).mean().mean(axis=1).dropna()  
global_total_cases_df['volatility'] = global_total_cases_df.rolling(2).std().mean(axis=1)  
global_total_cases_df = global_total_cases_df.dropna()  

In terms of correlation a simple plot of the total cases reported in each areas show that all countries are following similar patterns that 'could' be an interesting route for mapping pandemic pattern of disease spread.

Hypothesis:

SPY and SHY z-score-normalized minutely dollar volume difference plotted as a heat map can be used to detect COVID-19 global pandemic related shifts in market regimes

Discussion:

Extreme relative dollar volume movements in SPY and SHY, both intraday and near end-of-day, indicate a COVID-19 pandemic related shift in the market regime, relative to the earlier period of 2020, when the economic impact of the pandemic had not yet been factored into the market. The data also suggest that U.S. government market interventions may play a role in driving changes in market regimes (further investigation is required to reveal explicit correlations, e.g. passage of a major economic stimulus package by Congress, and Federal Reserve actions).

@Grant: That's a really neat idea. Two thoughts:

• The Jet colormap is usually frowned upon, see here: https://jakevdp.github.io/blog/2014/10/16/how-bad-is-your-colormap/
  
• There are definitely clusterings visible in the data. I wonder how much would be lost if you averaged over days. That would then allow you to more clearly link this to the spread of the pandemic which is data currently not included in your analysis.
  

Thanks Thomas -

I've attached an update. I am using a 1-year baseline window for the mean and standard deviation, and a 1-day window for the dollar volume smoothing. The latter window could be changed for better intraday resolution (see code pasted below).

As far as linking to the pandemic spread, my read is that the Fed did something right that took effect the last week of March 2020 to fix the market plumbing (see https://www.federalreserve.gov/covid-19.htm), versus a glimmer of hope that the epidemic was in check, and that the economic impact was understood and priced into the market.

Here's an early article (published Jan 27, 2020,10:20am EST) highlighting the black swan event risk of COVID-19:

https://www.forbes.com/sites/jackkelly/2020/01/27/the-coronavirus-is-a-black-swan-event-that-may-have-serious-repercussions-for-the-us-economy-and-job-market/

The question in my mind is how many other similar public statements are out there, and is there any indication in the market data that actions were taken in response to them (e.g. large moves closing out long stock positions)? Does Quantopian have access to ETF flow data?

WINDOW_BASELINE = 260*390  
WINDOW = 390  
dollar_volumes = data.iloc[:, 0:2]  
means = dollar_volumes.apply(lambda x: pd.rolling_mean(x, window=WINDOW_BASELINE))  
sds = dollar_volumes.apply(lambda x: pd.rolling_std(x, window=WINDOW_BASELINE))  
dv = dollar_volumes.apply(lambda x: pd.rolling_mean(x,window=WINDOW))  
zs = ((dv - means) / sds)  
data['z_diff'] = zs.iloc[:, 1] - zs.iloc[:, 0] # DV_SHY - DV_SPY (where DV is the z-score normalized dollar volume)  

Want to submit to our COVID-19 challenge, but don’t know where to start? Check out our latest video, “COVID-19 Infographics Challenge Overview” to get an in-depth explanation of the challenge and the possible approaches you could take: