11
$\begingroup$

I have been hacking around with algorithmic trading as a hobby project to build my data analysis skills, coding skills, and learn more about financial markets. As part of this project I am interested in developing a web interface to monitor my portfolio. I have done some homework but would like to hear the community's thoughts, as there seems to be much more focus on backtesting and other analysis in the open-source community (for good reason).

I am hoping to build something (in python) that could be used to monitor a real-money portfolio, as well as be used as a project on my CV that would not be laughed out of a reasonably sophisticated trading shop. My general timeframe per trade will be 5 minutes or longer, so nothing most folks would refer to as HFT.

What features* should an investment-grade algo dashboard include? Are there any examples of note?

Clarification: I am not just looking for visualizations, but also standard/key metrics. For instance, I would expect to see some measures of value at risk, returns over key periods, etc.

$\endgroup$
  • 8
    $\begingroup$ I suggest a Kill Switch $\endgroup$ – Alex C Feb 17 '16 at 3:26
  • $\begingroup$ do you have your own money invested in it? This will be one of the first questions you're asked - if your answer is no, then the very easy follow up by them is "why not?" which you obviously then flounder with. If you are investing in your strategy, then it's a different game, but then I suspect you would be less likely to be asking this question...if you're merely creating an analytics platform, then without experience, how can you know what is required by users? $\endgroup$ – will Jun 4 '17 at 21:43
  • $\begingroup$ I have my own money invested in a number of strategies, but this question is asking about generic KPIs and summary statistics that traders and portfolio managers want to see independent of the underlying strategy. $\endgroup$ – altfund Jun 8 '17 at 18:58
  • $\begingroup$ @altfund Did you end up building your own custom dashboard? I'm currently working on one myself for crypto.. $\endgroup$ – Goose Apr 28 at 3:15
7
$\begingroup$

Unfortunately, the answer is: it depends. People care about different metrics and visualizations depending on the type of strategy that they are running. It is a very bad idea to spend time creating visualizations without knowing what you are using those visualizations for.

A common feature is a 'table-oriented' layout of data about your orders and executions. This is the lowest-level information on which most of your other metrics and visualizations are derived. Every commercial trading system with a UI that I've come across has this.

You should play around with a few free trial or open source trading systems and get ideas from there.

$\endgroup$
4
$\begingroup$

Believe me, you won't find anything out there. I've been in the same position as you several years ago. So, I decided to design the dashboard myself. Below a link of what I did (html5)

trading dashboard

$\endgroup$
0
$\begingroup$

Pick a strategy that "works" for the given time period. Maybe correlation/pairs trading? Bare minimum here are intraday levels, intraday divergence, price ratio, correlations for different time periods, standard deviation (or some other such measure) bands.

Keep in mind you don't need to necessarily "buy in" to the strategy for a first version. Don't spend too much time on this part (for now) and focus on the underlying mechanics. Once you have a working dashboard for a given strategy, you'll find the work for adapting it to some other strategy is a lot easier.

You'll spend way more time than you think on "plumbing" and infrastructure, so your goal should be to take the current enthusiasm you have and use it to get to a finished/usable product as quickly as possible. Good luck!

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.