Time for an update
Whilst we’re still early in the new year, the ideas I had to refresh the ETFtracker app were still fresh in my head.
As a reminder, we have 4 key areas of the app:
Single ETF view
Analytics lets you look at the various metrics produced by ASX and Chi-X historically in an interactive way across different sub-menus. Single ETF View pulls multiple metrics together onto one page to give you an overall ETF snapshot. Data Stories is a semi-interactive look at some key stories coming out from the ETF data each month.
The Dashboard section needed an overhaul though. It showed some bar charts with rankings of spreads and net inflows. It had a scatter chart for 1-month vs 1-year price returns. It also had 2x line charts showing number of trades each month and the number of Funds/ETFs listed on both exchanges.
Whilst there was some filtering that users could do the big limitation was that, for a dashboard, it was not really a dashboard.
Putting our data analytics hat on for a second we know that dashboards are meant to convey some simple key messages and display them in an easily digestable way. The best dashboards contain data and charts that need little to no explanation.
For ETFtracker to be even more useful and insightful it needs some help.
Enter the heatmaps
So the solution to this was to look to the past and what I used to produce when I was doing my macroeconomic dashboard at QMG (QuantMetriks Global).
One of the key mainstays of our monthly analysis there was a heatmap. It was an easy way to use colours to showcase the different industries and sectors we covered and what was doing good or bad in terms of metrics like Sales, Margins, Price, Volumes and more.
Heatmaps are great as they are commonly seen and understood in finance but they have their limits. For example, looking at the following heatmap from Finviz.com, we can see that they only give a snapshot at a point in time of what is going good and bad across a basket of stocks, sectors or industries.
They only tell you what is going on right now without conveying the message about what has happened in the past (or if you forecast, what could happen in the future).
Bar chart with trendline
One suggestion is to include a bar chart which ranks what is good and bad along with an accompanying time series chart that reflects the selection.
This is good but does require good labelling and for users to click on each bar to highlight the related time series.
Small Multiples / Multiple trendlines
Another option is to do a small multiples style chart which shows major categories broken out as a number of individual time series charts.
When done with some colour, this is a great way showcase what’s happening now as well as what the trend has been in the past.
Experimenting with our own updates
Putting all this together, I am currently experimenting with a dashboard revamp using a mix of the best of both worlds. The colour from the heatmaps to show a snapshot of what is good and bad and the small multiples approach to showcase trends.
There are a number of ways we could do this but we’ll start by looking at the different categories that the exchanges define. There are a lot of them so I’ve grouped similar ones together like the various domestic equity sectors. They just sit under the banner of Equity – Australia. All the other global equity related ones are grouped under Equity – Global. Doing this got us down to 13 categories so to get an even number we’ve cut the various Index related ETFs.
The trends on their own only show part of the story. We needed some metrics as well as to say whether or not the latest month is good or bad. Simple background shading shows whether the metric for a particular ETF category is good or bad.
This change isn’t deployed into the app just yet but here’s how things like for January 2021 for a measure like Net Inflows.
In the following chart, we can see a sea of green as all categories had positive overall net inflows even though going from December 2020 to January 2021 there were some declining trends.
Contrasting this with the same view in March 2020 and we get a different picture. With the market rife with uncertainty we saw a more outflows than inflows and more red measures in the mix.
There are some areas to improve including ensuring a similar Y-axis across the trends otherwise different categories are not easily comparable. This works with total price returns for ETFs (see below) but with the ASX:MGOC listing in November 2020 it skews all the other categories if we align their Y-axis. So, there's still some thinking and work to do before deploying them.
Anyway, what do you think about all this? Do you like this idea of displaying the data and do you see ways to use it even in your own data analysis?
Stay tuned here for more ETF market updates.
PS: the Finviz.com site is awesome as it does display some pretty cool interactivity as you hover over each heatmap element. It interactively shows you consittuents and trends so they've done well with that and it is something that can be done with ETFtracker too but we'll save that for a later date.