Skip to content
All library documents

Designing an Algorithmic Trading Dashboard Around Strategy Needs

Article Quant Q&A · Author: altfund

Summary

The discussion considers what an interface for monitoring an algorithmic portfolio should show. Its central design principle is to choose metrics and visualizations according to the strategy and the decisions the dashboard must support, rather than building charts without a defined purpose. A table of orders and executions is proposed as a useful foundation because these records underlie many higher-level measures. The question also mentions risk measures and returns over selected periods as examples of desired metrics.

For a pairs-trading example, one answer suggests monitoring intraday price levels, divergence, price ratios, correlations across periods, and dispersion bands. The replies recommend examining existing trading systems and building an initial dashboard around one strategy before expanding it. The discussion offers no tested dashboard specification, comparison of products, or evidence that particular measures are universally suitable. Its suggestions are starting points; the appropriate display depends on the instruments, strategy, and monitoring purpose. It also cautions that data handling and infrastructure can take substantial effort.

Key ideas

  • Dashboard metrics should be selected to support the strategy and decisions being monitored.
  • Order and execution records can provide a foundation for higher-level portfolio measures.
  • A pairs-trading view may include price divergence, ratios, correlations, and dispersion bands.
  • A first dashboard can focus on one strategy and expand as requirements become clearer.

Tags

Full text
# Which features to include in an algorithmic trading dashboard?


# Which features to include in an algorithmic trading dashboard?












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.

## Answer by madilyn (score 7)

https://quant.stackexchange.com/a/24347

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.

## Answer by Ariel Silahian (score 4)

https://quant.stackexchange.com/a/24379

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

## Answer by Todd Page (score 0)

https://quant.stackexchange.com/a/24362

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!

Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.