Writing a story with alt data

I seem to spend most of my time writing. For over a year Alexander Denev and I have been coauthoring The Book of Alternative Data. I’ve also written a few papers on alternative data and FX recently. Aside from that, I’m still writing, although usually in Python and I’ve also recently started to learn kdb+/q. The syntax of q is pretty different to Python and it’s also a functional language. It is however ideal for dealing with tick data, which is the main reason I’m learning it. I’m not sure anyone would ever contemplate writing a story in Python or any other computer language for that matter, and whilst it is in English, our new book, can hardly be described as a novel, even if there is an underlying plot line, namely that investors should use alternative data! However, all this writing and let’s call it pseudo-writing, when it comes to coding, did get me thinking about the whole concept of storytelling and how it isn’t purely in the realm of literature.


Ultimately, a lot of what we do is about storytelling, outside of writing. If we’re an investor scouring the market for ideas, we want to find a story that we can buy into. If there’s no story we can articulate about why we want to invest, then the investment decision seems questionable. If we’re a quant, we seek to find a story, essentially a hypothesis, that can rationalise why we’re building a certain model. If there isn’t a story, the plot line invariably involves that villain of systematic trading: data mining. Often in financial markets, a time series provides our basis for a story, that can be explained in a visual form.


Alternative data provides us a richer vocabulary to tell our stories in financial markets. Ultimately, we will still usually distil these datasets into time series for consumption. However, our sources are often more disparate and can include images and text. We can also visualise datasets across different dimensions and innovative ways, a subject I discussed recently in another article. If we have geospatial information, such as the location of an event, we can display often very complex datasets succinctly on a map, something that has been utilised extensively during the current COVID-19 pandemic. We can combine location information with animation to provide a view of them changing over time.


We might not use the lavish vocabulary of Shakespeare or Dickens in financial reports (and somehow I’m not sure whether traders would read such lengthy work, where brevity is so prized). However, this doesn’t mean that we shouldn’t use alternative data to imbue our stories with more description. In an increasingly data dependent world, it is important to use the tools at hand to improve our understanding of what’s going on. It isn’t about dispensing with traditional datasets, it is more about augmenting them with alternative datasets. If you’re ignoring alternative data, you’re ignoring a lot of the elements of the story, and worse than that, you might end up hearing the story, once it’s old news to other market participants.


Using alternative data isn’t easy, as we shall describe in our forthcoming book, however, there are many steps you can take to make sure that your investment process can consume and interpret alternative data, to write better stories and hence most importantly make better investment decisions. If you’re interested in using alternative data, take a look at our new book. I’ve also created a Python & alternative data workshop which I can teach at your firm (or via video conference) to help introduce the subject.