Finding the right investment

Creating and managing a mutual fund portfolio can be quite a task without the right set of tools to compare funds. On the one hand, you have volumes of information from research reports. On the other hand, you have data on Excel that let you compare specific parameters. But few tools offer a complete view of the portfolio with all relevant details across the funds in a single snapshot.

At Gramener we applied visualisation to a set of equity funds, comparing them on various parameters.

The layout has been kept simple, but the information density is rather high. Both are on purpose, This, clearly, is quite a complex visualisation and merits closer observation.

The first block shows the name of the fund, the current net asset value (NAV) and the overall trend over the last 30 days. The black line shows the daily trend while the light blue line behind it shows the movement of the benchmark – the Sensex in this case.

The next block shows daily returns as a horizon graph. The returns on each day over the last 30 days (from left to right) are shown in red (for negative returns) or green (for positive returns). The intensity of the colour indicates the magnitude of the return. This makes it easy to compare returns for a given period across funds. For instance, it is clear that the beginning of the month was a bad time (since the left-end is red) for almost every equity fund.

The returns are also plotted against the returns of the Sensex as a jitter plot. The red dots on the left are days when the returns fell below the Sensex. Far left indicates very low returns, far right indicates very high returns. A glance shows that the second and third funds have returns that are spread out, but the others are fairly closely clustered around the Sensex’s returns.

The next block shows the average return over longer periods – a week, month and quarter. The colour indicates how high or low the returns are with respect to other funds. Red indicates the lowest return, green indicates the highest return. Here, the second and third fund have had a fairly high return compared to others during the last week.

Some of the more complex parameters that investors look at are Jensen’s alpha, beta, sigma and the information ratio. These too are included in this visualisation, again showing their relative performance with respect to the other funds. The only fund that does not perform poorly on these parameters here is the third fund.

But one other parameters that would be of interest is, how good is the fund’s return compared to other funds. Irrespective of the return with respect to the Sensex, a key parameter is the return with respect to other equity funds if these funds are the investments of concern to us. We plotted the weekly return as a percentile rank. If the fund is the best performing fund on a given day, we plot a dot at the right end. If it is the worst performing fund, we plot a dot on the left end. The mix of dots tells us how the fund has performed relatively over the last month.

The last column shows the average of the weekly percentile rank. This is a good indication of the fund’s relative performance averaged over time. The first fund, for instance, has on average outperformed 63.7% of the other funds. The next has outperformed 61.6% of the other funds. And so on. The table is sorted based on this column.

The aim of this visualisation is to present, in a very condensed and information-dense fashion, all the information an investor requires to make a decision. As you can see, an explanation of the visualisation takes more space than the visualisation itself! Yet with a bit of understanding, getting insights out of this plot can take just a few seconds. A far cry from having to read 50 equity research reports!

When to invest

Sometimes, timing is everything in investments.

Last year, The New York Times published a piece titled In Investing, It’s When You Start And When You Finish. This showed the significant impact of timing in investment decisions.

At Gramener, we applied the same visualisation to a few Indian stocks over the last 5 years.

Here’s what it looks like for ICICI’s stock.


If you invested in ICICI stock in Jan 2007, the first row of boxes show the kind of returns you would have seen.

The colours indicate the degree of profit or loss. Red for losses, green for profits, and yellow for neutral values. Selling in March 2007 would have made significant losses. Selling in Jan 2008, one year later, would have given you a good profit. And so on.

The same is extended to investments made in other months.

The black boxes show a holding pattern of 1 year, 2 years, etc. You can get a sense of what kind of returns you would make with a strategy of holding for 1 year, 2 years, and so on.

Here are similar pictures for Infosys stock and SBI stock.

At Gramener, we took a look at a number of such stocks and their performance over the last five years. A interactive app showcasing sample of those is available at

Visualising securities correlation

If you were wondering how the securities in the world move against each other, the picture below is the answer.


This picture shows the correlation various currencies, indices and commodity prices by linking together three powerful types of visualisations.


The first is a coloured correlation matrix. In this picture, we have three securities: the British Pound (GBP), Gold Price (XAU) and the Dow Jones Index (^DJI). The price of GBP and Gold tend to move slightly together, and have a correlation of 0.36 (36%). So the cell that’s between GBP and XAU is marked with 36. Similarly, the cell that’s between XAU and ^DJI has a –64 because Gold and the Dow Jones index are slightly negatively correlated.

The colour coding is based on the correlation. Red is -1, Green is +1 and Yellow is 0.

securities-correlation-scatterplotThe second is a scatterplot matrix. The cells that mirror the correlations have a series of dots. Each dot represents the price on a particular day.

For example, compare Gold and the Dow Jones. It isn’t a straight-forward negative correlation. In fact, it almost looks like there were two periods: one in which gold was high when the Dow Jones was low, and vice versa. But within those periods, there appears to have been a mild positive correlation.


The third is the hierarchical cluster. The securities is grouped into similar ones based on their correlation. For example, GBP and Silver (XAG) and reasonably close to each other, and form one group. This group is most closely related to the Euro (EUR), and the three of them are closest to the Australian Dollar.

Arranging the securities by the hierarchy makes it easy to spot groups of securities that tend to move together.

For example, in the original visualisation, there appear to be a set of logical blocks


At the centre, four securities – the Pakistani Rupee (PKR), the Sensex (^BSES), the FTSE (^FTSE) and the S&P (^GSPC) – tend to move together, with each other; but move in the opposite direction to the next group of securities – the Singapore Dollar (SGD), the Japanese Yen (JPY), Gold (XAU), the Swiss Franc (CHF) and the Chinese Yuan (CNY).

Similarly, the Swedish Krona, Canadian Dollar, Indian Rupee, Hong Kong Dollar and Mexican Peso form yet another group that moves together, but in the opposite direction from the strong Asian currencies in the block above.

We at Gramener have named this visualisation a clusterplot. It’s a powerful technique when applied to time series of multiple (typically 5 – 50) variables.

Here are some cases you might consider using them:

  • Group your products based on consumer behaviour. Which products tend to sell together? Which ones cannibalise the sale of the other? Is there a way of rationalising the product base to reduce complexity – without losing customers?
  • Group retailers based on sales. Which retailers tend to cannibalise the sales across each other? Which ones complement each other? Where would you need to rationalise to avoid duplication or overlap?
  • Analyse process quality drivers. For example, if temperature, pressure and salinity affect your product quality, what impact will increasing one parameter have on the other?