Visualizing Financial data as Scatter Plot Matrix- An Alternative approach

Introduction:

Scatter plot is a widely used graph/plot to visualize data especially when data is related to fields such as Finance and Economics. It has been widely used to represent relationship between two sets of data. Financial analysts often use these plots to show correlation between two funds, two stocks or even a stock and the general market.

So how do you represent data when there are more than 2 variables? Let’s say a portfolio of stocks or ETF or Foreign Exchange or Commodities, or Corporate Bonds. Now we are in a territory of multiple variables. This is important as we would like to study their correlations. There are a few ways to represent this data and one of them is scatter plot matrix. The scatter plot matrix is not a very new concept and has been used in the fields of Medicine and Biostatistics. The implementation of this visualization technique is new to finance and not widely used.

Interpreting the Visualization:

If we observe the plot it looks like a square matrix with rows and columns. The labels are on the diagonal and the returns are points in the matrix (off diagonal). For time being we will only look at the image below the diagonal. The best way to interpret the plot is to go down the row and look on the right. For example, the second row and first column displays the scatter plot of goldman and citi whereas the third row and first column displays the relationship in returns of goldman and apple.

Following similar rule, to view the relaionship between facebook and citi we will observe the image under fourth row and second column. The section of the plot above the diagonal (upper diagonal) is a reflection of the image below the diagonal. Hence, if  we wish to reduce the complexity we can remove the upper diagonal. The visualization below looks much tidier in absence of upper panel.

Note :

1) It is not possible to generate the scatter plot matrix in MS - Excel.

2) Due to the restrictions on the size of images in Linkedin the plots look stretched.

References:

• JunkCharts : Provides another application of Scatter Plot Matrix 
• Myblog : I try to post some new ways to visualize quantitative and qualitative data
• data - The data for reproducing the plots.save the file as a .csv extension to avoid any issues.
• code - The file consist of code to reproduce the 3 plots along with the explanation about each function used.