Learning coding as a fundamentals analyst?
Hi everyone,
I am someone going to a target school interested in working at a fundamentals hedge fund. Seeing how quantitative and automated the hedge fund industry has become, is it worth it for someone who's more interested in the fundamentals to learn to code languages like Python to understand the quant side of trading? Or is it better spending my time getting specialized and better at fundamental analysis? I am reading that macro/fundamental funds like Bridgewater have moved into a "quantamental" model combing both quantitative and fundamental analysis. I am also interested in the alternative data side of finance.
Sure, learning "coding" is a very smart choice that will definitely help you in the future. It will also develop your strategic thinking when you move from tutorials to more advanced programming.
I suggest you to learn Python & C (+ C#). They will definitely help you. Good luck and have fun while learning!
Thanks! What languages and concepts should I be familiar with if I want to go into Alternative data?
For python - pandas/numpy and working with arrays, for practicality you’ll want to be able to scrape (beautifulsoup) and work with reading/writing to excel. This is from the perspective of a traditional fundamental analyst using coding to work with alt data and increase efficiency. The answers are probably different for other applications and depth of focus. I think it’s best to think of a representative task and learn around that using online tutorials. Railroads and airlines are some examples of datasets that might be good to practice with as they’re standardized and free in the US.
Gotcha. So as a fundamentals analyst, coding is useful to extract and visualize and model data, rather than make algos or discover patterns to trade.
Correct, as a traditional fundamental analyst it is most useful to collect and work data and to automate rote parts of the work process. I’m not versed in ML - if I was I can think of helpful ways one might be able to search for and draw insights from large datasets or their own internally generated trading/portfolio data. But my time is probably better spend finding the next investment idea and then using basic python as needed to diligence that idea, create and analyze datasets to track KPIs. The longer term your investment mandate is, arguably the lower return-on-time programming is.
Do you think it is worthwhile then as a fundamentalist to learn machine learning? And since so much of our work revolves around Excel, should I learn VBA and R too?
I don’t really know, but in the sectors/strategies I’ve dealt with it MAY have been incrementally helpful, but probably not as helpful as finding the next idea or expanding your coverage. To the other poster - yes VBA would also be helpful in general and to some degree with alt data. The better you are at making excel work as an input/output for your scripts, the more I think real value add you can be on a traditional fundamental equity team where the PM probably isn’t interested in migrating the financial models completely out of excel or anything like that. To that end I’ve found SQL to be helpful for large datasets and as a complement to python and Excel. Other people with more coding experience might certainly disagree - but I used SQL and python to help me when tasks were too large or too slow for excel - not necessarily too mathematically complex for excel which isn’t really an area I’ve dealt with. In my experience in a fundamental role, big tasks that are really complex should be outsourced to the data team at your fund or a consultant in most cases.
Wow thank you! Other than modeling, do you use your code to execute trades or make trade alerts? What do you think of web-scraping sites like Stocktwits to determine market sentiment and base trade off of that?
For advanced modeling, what are some maths I should learn?
I don’t use code for actual trading, though I have an excel for my portfolio that generates a string that I paste into an email to out trader or into the firm system.
Social media / qualitative sentiment monitoring seems to be one of the most commercialized / commoditized usage of alt data and I’ve never found it worth paying for personally.
From my experience as a fundamental analyst - learn all you can about your companies and coverage, find the key questions that need answers and the key KPIs that need forecasting. THEN explore what data might be available and how you might be able to collect, evaluate and use it. I think starting with the data and trying to find an application for it - unless you’re at a company that is generating substantial exhaust data in the first place - is a less useful and you see a lot of people marketing alt data but misunderstanding the potential usefulness and utility of the data for my purposes. And that’s annoying - I want to buy the raw data and use it how I want to use it. I have no idea what kind of backward looking smoothing bs you might me using etc. I just want the ingredients - and I don’t want to teach a vendor how to make the pie cause I don’t want them to commoditize and market it. This misunderstanding between vendors and myself is why I don’t pay for 90% of the data that is marketed to me even if it might be useful. If I can’t vet it, it’s drastically less actionable. And if you market the final answer (credit card data) fine I’ll pay if everyone else is using it and I have to but I’m not excited about it.
Eos dolores commodi accusamus fugiat perspiciatis. Exercitationem sint qui eum qui provident qui. Vel dolor ipsa est cumque.
See All Comments - 100% Free
WSO depends on everyone being able to pitch in when they know something. Unlock with your email and get bonus: 6 financial modeling lessons free ($199 value)
or Unlock with your social account...