Web UI for non-python audience

[nrpalao]

Predict crypto trades with machine learning right from your browser

Posted my idea in the Slack group and started working with @jchomali on a web version to bring all the potential capabilities of the trader to people not familiarized with python or non developers.

The main intention is both:

• To create finally a good looking, clear and easy to use dashboard to see the situation of all your investments using Poloniex API
• To receive suggestions argued by machine learning about your next movements.

Main view with a quick recap of all your investments and main situation

Detailed view with statistics about each one of the coins where you invested and some more graphs about your detailed activity

Just a quick draft, lots of improvements to be added but that’s the main idea. Would love to receive your feedback, thoughts and if someone wants to work on it, just ping us!

[NDuma]

"I like the simplicity of the Circle & the Hover expansion. Would love something like that for navigating current positions & indicating overall portfolio allocations."

https://pytrader.slack.com/messages/gui/

[owocki]

Hey @nrpalao I'm excited about this. If anyone has browsed the /admin of this repo, they've seen how poor my design skills are .

For reference, here are the UI views presently in the repo

1. Django admin
2. Portfolio view
3. Trader optimizer
4. NN optimizer
5. Classifier optimizer

An important part of 4 and 5 is to be able to drill down on different parameters (for example, in this screenshot I can see that Naive Bayes is a better algorithm than many of the others, so I can click on it and then on the next page the data is constrained to be only Naive Bayes, and I can see the results broken down by other parameters. I use this UI to optimize the system (more documentation coming on this today).

Some mock ups I would like to see:

1. A UI for tuning neural networks and classifiers
2. A UI for gauging the success (or lack thereof) of the trading engine.

[NDuma]

I got some licenses for Balsamiq mockups on another open source project; however, I'm not sure if this one is large enough yet; or, if it is a good solution for collaboration here.

Would be willing to check into it if there's interest.

What is the plan for the WebUI repo? As pyTrader is not under an Org. Can we get this linked up in the ReadME and also a link here?

[nrpalao]

@NDuma ok, great feedback ->

• Added a graph with all portfolio allocations (so useful indeed): hover to see initial investment and next to the coin name you see the fluctuation and actual price
• Moved previous box content to the first block so you now can see in a first view: total value and money won + growth selected

Full view

[nrpalao]

Hey @owocki, thanks for your comments. No worries about the current design, we will make it as easy and sexy as possible to trade and keep track of your portfolio, obviously taking advantage of our powerful backend!

Replying to your proposal, IMO (and please correct me otherwise), the current interface is difficult to be interpreted and some information is duplicated. The core idea behind my interpretation of the project is that anyone, without special technical or trading knowledge can use our tool, to manage its portfolio, while our algorithm studies which paths you can take for a notable revenue increase, with at less human action as possible.

Next steps:

In that point of the project, I would really like to understand which graphs of the actual system you think that are really needed and must be added to the web dashboard and which inputs the user can apply to the trader (in addition to stop losses or sell limits) in order to finish defining all features.

Comments:

Some mock ups I would like to see:

1. A UI for tuning neural networks and classifiers

Can you please elaborate more context or share an screenshot about how that works? We need to keep that kind of adjustments as easy to be understandable and to be taken as possible. Also, as much decisions we can take by the user, better IMO.

1. A UI for gauging the success (or lack thereof) of the trading engine.

Currently you can see the % of good and bad decisions taken in a simple graph (that probably needs some more adjustments).

[phon3]

Hello, I was wondering if I could get an invite for the slack channel to join in on the discussion at pytrader.slack.com? Sorry if this is not the place to ask.