Market risk management reports often contain many ‘Top N’ tables. Figure 1 shows a typical example. Each of the four tables shows the top 5 underliers (companies) for a different risk factor – in this example delta, gamma, vega and time-weighted vega.
Figure 1 actually underestimates the prevalence of these ‘Top N’ tables. Often there may be about for about 9-10 ‘Top 10’ tables and even more if underliers are split between indexes and individual stocks.
It is not easy for readers of these tables to gain insight about the data in these tables quickly . Firstly, they suffer for all the usual problems of tables of text and numbers. Secondly, it’s quite hard to follow the details for a particular underlier across all risk factors. Say we are interested in ‘Underlier F’ since it has the highest delta. We have to scan the other three tables to find it and in fact we don’t find it in the gamma and vega tables since it is not in the Top 5 in these risk factors. Thirdly, the tables don’t provide that much data – there are only 20 data points (4 tables of 5 numbers each).
There are a few alternatives that may improve on this. We can add data-bars that provide some visual cues about the relative values as in Figure 2.
However it would be much better to see all the data for each and every underlier that appears in any of the tables. In our example, that will be between 5 and fewer than 20 underliers since we can expect the same underlier to be present in several tables. Figure 3 shows the data organised with each underlier on a separate row. Now it is easy to see all the values for a particular underlier. Although it is now harder to determine the biggest value since we are no longer ordering by value, the data-bars help us quickly home in on the important values.
The data points represent a snapshot of the position as of close of the previous business day and it is very useful to know the recent history – how has this value been arrived at. In Figure 4, sparklines were added to show a micro-chart of the each value over the last 10 days.
Figure 4 has made good use of precious space and increased the density of data points. As mentioned Figure 1 represents 20 data points. Figure 4 however manages to represent twenty times as many data points (i.e. 400 data points; 10 underliers by four risk factors by 10 days) while only taking up about twice the screen or page real-estate.