Data visualization and analysis using the relevant historical background Under-19 COVID the US stock market crash

If you do not know the US stock market fell in the past two months, then you are either an ivory tower university students, or both is no investment and no student loans to repay a lifetime of low-level workers. In any case, regardless of whether you have not put a lot of family property in the stock market, you should know what happened in the past few weeks. Whether you are lucky or unlucky in this crisis, you have witnessed a black swan event.

Tips: Investopedia black swan event is defined as:

Black Swan event is intended for those who exceeded expectations under normal circumstances, and has potentially serious consequences of an event totally unpredictable. Features Black Swan events are extremely rare, far-reaching, and has been particularly prevalent in hindsight.

In this article, I will use the data and scientific means to quantify the impact of the decline in the US stock market in 2020, and compare it with the recent (and earlier) other major decline in history. I will collect data to help build my model, set up to provide comparative statistical background and to communicate important ideas through visualization.

One last thing, I only use publicly available data, all of the code so that the data used in this text visualization and raw data can be used on my GitHub project. Now, here we go!

Over the past few weeks to quantify the market turmoil

We can use a number of different data sets to understand the market turbulence. We first analyze the S & P 500 Index since the past three months (1 February 2020 to 20 March 2020). This is my data set looks like:

 

You may search the Internet every time the stock will see these quantitative indicators; now, let's do some analysis in order to better understand them. First, I'll show you a simple price - time curve, it has shown a lot of information.

 

You can clearly see that in about mid-February (February 19), the market began a sharp increase in volatility. It should be noted that the focus of our concern is the severity of the concussion of market volatility, rather than the trend is up or down. Let us further, additional to the time-varying market volume visualization, this indicator shows the number of ongoing transactions.

 

Now you can also see market volatility and volume are increasing, particularly in the last few weeks. On the surface at least, another thing you can see from the image above is the first real surge in trading volume a day or two after the experience of the first few in the market crash.

Let's use another visual way to see the percentage change in the price and market volume. The percentage change index using different scales, so we need to generate a further description from the raw data. A good method is to use the Z-score. Briefly, Z-score is a data point in a display of how far from the mean statistical indicators. Given our sample size over the past three months, the average deviation is great, so I will use a modified Z-score, which depends on the median, because it is less sensitive to outliers and data skewness . Also, I will only use the absolute value of Z-score is modified, because I only care about the value of distance, not direction. You can see the results below.

 

We have now confirmed that the percentage of abnormal changes in the first volume itself abnormal. Can be seen from the figure, something happened February 22 weekend, resulting in the percent change from the standard deviation rose to 0.5 ~ 2 intermediate values ​​to standard deviation. In a follow-up article, I will explore the coronavirus cases that weekend, to learn more about what happened.

Currently we use the Z-score point of view it is useful to visualize data, but the following visualization further plans to capture the whole market panic spread, and I simply draw the daily percentage change in the S & P 500 Index. Note that in the past few weeks, the peaks and valleys of how the explosion, and the number of block diagram of the same data outliers.

 

We spent the last few minutes of analysis of US stocks fell alone. In the next section, we will put a drop fell on comparison with other recent history to get a better historical background, and really answer the following questions:

The market fell How bad?

Awful awful.

At any time we have to understand market behavior, one of the most useful strategy is to get some historical background. In this section, we will compare the US stock market fell in 2020 and other recent historical data, further analysis of market reaction.

 

To this end, I have collected the last 10 stock market fell more than 10% of the data. As you can see, the list recorded about 10-20% of the market decline, as well as 2007-- 2008 financial crisis and the dot-com bubble in 2000, which is much more serious even than the two crises. After that, I collected the S & P 500 daily rate of return within this time. My goal is to combine these statistics, the rate of decline in the stock market in 2020 and the rate of decline over the last 20 years were compared.

After the idea of ​​a number of different data visualization, I decided to use a line chart to best demonstrate the seriousness of the most recent decline. Visualization can be seen below.

 

As you can see, I constructed a chart to show the sum of $ 10,000 decline in investment in the market down in the process value. So that we can be equal and specifically comparison of different markets fall on history. As can be seen from the figure, the US market in 2020 fell much more serious not only in absolute terms of investment value decline, but the rate of decline should be much faster. I will chart a different data separately, as shown below, you can also observe this in the list below.

 

Well, now we have done! 2020 US stocks fell by no means trivial. In 30 days, the total market capitalization of the S & P 500 index of evaporation of about $ 8 trillion. More specific, the market evaporate about 800 billion dollar.

Author: Harsh Rana

deephub translation Group: Alexander Zhao

Watch deephub-imba send 20200405 to get code for this article github address

Published 30 original articles · 87 won praise · views 80000 +

Guess you like

Origin blog.csdn.net/m0_46510245/article/details/105333550