QuantStart Content Survey 2020

In this article we look at the recent results of the QuantStart Content Survey For 2020 newsletter. We look at which areas were the most popular as voted by the QuantStart community and we outline some of the topics we will be writing about in the upcoming year.

Back in 2017 we ran an upcoming content survey to find the topics that the QuantStart community were most interested in learning about. The results were very interesting, with Advanced Machine Learning & Deep Learning taking the top spot as the most popular discussion topic.

Recently however the QuantStart team internally agreed that the interests of the community may have shifted somewhat in the three years since carrying out the initial survey. Hence we decided it was time to run another survey, with a slight modification of topics in order to see what might have changed.

This week we sent out an email to the QuantStart mailing list asking for which topics would be most valuable to read about in 2020. We received a substantial number of responses, many of which contained fantastic suggestions on what to write about.

We would first like to thank everybody who took time to respond. Tallying up the votes has been extremely valuable and will help us prepare the most optimal content for the community going forward.

The Survey

The email asked which of a set of seven topics would be of most interest. We made sure to keep them reasonably similar to the previous survey in 2017. However, we replaced Equity Strategies/Factor Analysis, Algo Trading as a Business and Alternative Asset Classes with Tactical Asset Allocation, Careers Advice and Cryptocurrencies.

We have been discussing Tactical Asset Allocation recently on the site. We wished to gauge how much appetite there was for more content on the topic.

In addition one of the biggest changes since the previous content survey has been the rise in cryptocurrency trading. We also wanted to know if there was a strong demand for information on systematically trading this asset class that was not being met with the current set of articles.

The following choices were offered:

  • Advanced Machine Learning & Deep Learning - Systematic trading with Machine Learning, Deep Learning and Reinforcement Learning based methods. How to install and use TensorFlow, Keras and PyTorch to setup a modern systematic trading research environment.
  • Cryptocurrency Trading - Systematic trading of Bitcoin and other cryptocurrencies, using trend-following, mean-reversion and Machine Learning-based methods. Tutorials on connecting to crypto exchanges via Python and open source libraries.
  • Tactical Asset Allocation - Longer-term investing, primarily with ETFs, using momentum/trend-following approaches. Classical and modern portfolio construction, risk management as well as equity factor analysis.
  • Coding and Data Science - Advanced tutorials on coding, software development, data structures and modern data science topics using Python libraries such as JupyterLab, Pandas, Statsmodels and Scikit-Learn.
  • Mathematical Finance - Tutorials in mathematical topics such as Linear Algebra, Calculus and Probabilty along with Stochastic Calculus, Stochastic Processes, Bayesian Statistics and Time Series Analysis.
  • Trading Infrastructure - Connecting to brokerage APIs, placing orders, receiving and storing market data, getting started with remote servers (such as AWS, Azure), monitoring and automated trading.
  • Careers Advice - Guides on best university taught courses, Masters in Financial Engineering (MFE) courses, MOOCs, interview preparation and career progression.

Many of you provided multiple votes, which made it clear that some of the above topics were of equal interest. Many also took the opportunity to mention additional topics beyond those provided above and elaborated on why such a topic might be of interest to the community. It was great to receive such feedback.

The results of the voting are as follows:

As with the previous 2017 survey Advanced Machine Learning & Deep Learning is still the most popular topic, with just under a quarter of you voting for it. Given the popularity of machine learning and the rise of data science as a major career path this is still not surprising. What is perhaps noteworthy is that it does not have as clear a lead on the other topics as it did previously.

Once again some of you made it clear that you wanted to see tutorials of machine learning and deep learning applied to systematic trading specifically, rather than 'generic' deep learning tutorials on datasets that are of less relevance to quants. Hence our upcoming machine learning and deep learning content will now reflect this approach.

The next most popular topic was Mathematical Finance. It appears that there is a significant appetite for the core mathematical topics that underpin modern quantitative finance, including linear algebra, probability, statistics and stochastic calculus. We have modified our content roadmap accordingly. Hence in 2020 you should expect to see more articles discussing mathematics 'from the ground up', as well as how it is applied to quantitative finance.

Coding and Data Science was the third most voted topic. Once again this is not surprising given that software developers and data scientists are now rivalling quants in terms of being able to work with great companies, satisfaction with the role and lucrative compensation packages.

To some extent there is significant overlap with this topic and machine learning. When combining the votes of both it seems that over 40% of you wish to learn about programming, data science or machine learning in some fashion. Hence we will now start to expand QuantStart's content offering to include more diverse topics such as usage of core Python libraries as well as software development tools such as Git.

Tactical Asset Allocation and Trading Infrastructure were equally tied. It is worthwhile to note the rise of interest from the QuantStart community towards longer term strategies. Our new version of QSTrader, which is soon to be released, has been extensively tested internally on many TAA strategies. Hence there will be a significant amount of actionable content released this year on how to carry out your own TAA backtests. We will also be describing how to develop the necessary trading infrastructure required to actually run these backtests in an ongoing fashion.

Finally, Cryptocurrencies and Careers Advice round out the remainder of the votes. While cryptocurrency trading is a large part of the retail systematic trading space it seems there is less of an appetite for crypto specifics within the current QuantStart community. Although, of course this could simply be due to a statistical bias from the limited amount of crypto content that has previously been written on the site.

We are however surprised that reading about quant careers advice is not as popular as it has been in the past. This may be a reflection of how the quant finance field has become significantly more mature than when QuantStart was first released. Now that quantitative finance and systematic trading have a common educational path (undergraduate in mathematics, physics, engineering, computer science, followed by a Masters in Financial Engineering) it may be likely that less advice is required going forward.

Upcoming Content

With the results of the content survey email in mind let's take a look at what we'll be writing about in the upcoming year.

Advanced Machine Learning & Deep Learning

There are a wealth of useful tutorials on machine learning, deep learning and reinforcement learning available on the internet. However they often tend to utilise datasets where the samples are independent and identically distributed (i.i.d.), particularly from consumer web applications, possessing a strong signal-to-noise ratio.

Since quantitative finance and systematic trading make use of autocorrelated time-series, often with substantial noise, it is not obvious how to straightforwardly apply traditional machine learning and deep learning models to these datasets.

Hence much of the upcoming content this year will be spent discussing the difficulties of applying well-known ML/DL models to the types of data utilised in quant finance. In addition we will be highlighting how to build functioning signal generators (and thus trading algorithms) around these models.

Mathematical Finance

As mentioned above there is a significant appetite for learning mathematics for quantitative finance. The usual path to gaining these skills is to carry out a traditional taught course at university. However we have previously written on how to go about carrying out self-study in these topics without going down the university route.

While this categorisation is useful it can still be daunting for those undertaking self-study as many mathematics texts are not optimised for such an approach.

Hence we are currently preparing a set of articles on the 'key' topics and results needed for learning linear algebra, stochastic calculus and Bayesian statistics. These underpin the most common tools used within modern quantitative finance and should serve as a 'bridge' to help you understand more detailed, formal texts on the various subjects.

Coding and Data Science

As more and more individuals are learning to code, with many of those now choosing to become quantitative developers and data scientists, rather than pure quantitative researchers, it is necessary to learn how to use modern data science libraries and software tools.

Some members of the QuantStart team have a background in professional software development and are currently preparing articles on many of the methodologies that were once the domain of software engineering, but are now making their way into modern retail systematic trading workflows. Such tools and processes include Git version control, NoSQL datastores, MapReduce tools such as Hadoop and Spark as well as general proficiency with data structures, algorithms and the Linux command line.

These are essential skills for the modern quant developer (and even quant researcher!) and complement those skills already developed for data analysis, such as SQL, NumPy, Pandas and Scikit-Learn.

Tactical Asset Allocation

We have already begun writing about Tactical Asset Allocation (TAA) strategies on the site. It will be an often-revisited topic on QuantStart as we feel it is a great way to get started with systematic trading, without the difficulty of having to carry out trade execution on a daily basis.

TAA is also a useful 'gateway' into the topics of portfolio construction and risk management. These two topics are second nature to a practising institutional quant, but it has always been QuantStart's philosophy to expose retail traders to these concepts given how valuable they can be towards increasing returns and reducing losses.

We will be providing backtests and code implementations (utilising the latest QSTrader version) for many well-known TAA strategies in the near future, in the hope that you will be able to develop your own portfolios and dynamic asset allocation rules from these examples.

Another Big Thank You

As with our previous content survey in 2017 the QuantStart team would like to say a big thank you to everybody who voted and provided great feedback. It is always extremely valuable for us to know exactly what types of content you are interested in reading about.

If you have any further suggestions of topics that we have not discussed to date that you think would make great articles for the site then please feel free to contact us at support@quantstart.com with any suggestions you might have.