Name: 戦略メトリクスの可視化 - Ziplineチュートリアル Pythonによるローカルバックテストとファイナンス p.2 (Visualizing Strategy Metrics - Zipline Tutorial local backtesting and finance with Python p.2)
Uploaded: 2021-01-14T10:44:11.000Z
Duration: 9 min 10 s
Description: VoiceTubeの動画で発音を聞きながら英語表現を覚えよう！学べる英語：

分詞構句

And welcome to the second part of the Zip line local tutorial miniseries for back testing and doing finance with punny Thon In last tutorial, what we did was we just went through the arduous process of getting zip line running on our machine, running through a simple, just by apple strategy.

And we got to this point where we ran the strategy and we can see that it's making some output.

And no matter which way you run it, you got we at least said the output should go to Strat dot pickle.

So at least now, if you run it via the magic here, you do get the output saved to this apparent data frame creates this table that I'm telling you now is a data frame.

But but this pride isn't exactly what you were hoping for.

You probably want to see some sort of graph, and you also probably want to know what you're working with if we just use this scroll bar.

Um, at some point, I thought we would hit it.

Looks like we're actually seeing all the columns here, which is weird.

Yeah, There's other columns here that we're just not seeing.

So the first thing I'm gonna go ahead and Dio is, um basically, I've already run this, so I'm just filming from the last part.

So if you need to make sure you load the extension, run the algorithm again and all that, I'm just gonna continue one down here.

This did, at least in this the way that we ran it This time I still have one more way.

I'm gonna show you guys how to run things, but, uh, I just figure information over the overload.

Well, one option we have, uh, first of all would be like, let's go ahead and import pandas as PG.

Don't make it too crazy, but I want to be eligible.

Uh, and then let me go ahead and make some more imports.

Put this up towards the middle, so we're gonna import.

Let's import Matt plot lib dot pipeline as peel tea from Matt Plot Live.

Let's go ahead and import style, and then we'll do a style that use G plot because we all know in finance how beautiful your charts are Makes a huge difference in your performance.

So, uh, the next thing we're going to say is I'm just gonna say it back Testy f equals P d dot eyes It was it read pickle Freed pickle and then specify the name of the pickle.

In my case, we output to Strat dot pickle so Strat dot pickle So that reads in our data frame.

And then we could do something as simple as back test on her score.

D f dot uh, port folio underscore value dot plots guilty dot show boom.

You could also plot in line and then you don't have to call appeal t doubt show.

Ah, but I'm just always in the habit of doing it this way.

Turns out you do really well, if you could just buy 10 shares of Apple every single day without any respect to really anything at all.

And you just have infinite money anyway, uh, cool.

Now you might be like Well, Centex, slow down.

How did you know you could do portfolio value?

Well, one option was you could probably like, look up here and be like, um let's see if portfolio values even here.

I'm not seeing it, But a lot of times I'm just totally blind.

It might be there anyway, Uh, the way I knew about it is what you could say is, just now that we've loaded back in that, uh, this data frame and called it back testy f we could just dot columns this bad boy and boom.

So all of these things are things that we can graft right now they're automatically tracked, except for a p O.

We added that so, for example, if there's something else that you want to track in your strategy as it goes, besides all of these columns right here, if there is something else just like on quantum peon, you can use the record function and record it now.

In quantum peahen, there's a limit of five.

I think you could only record five things.

If Zip line locally is the same, I would have to imagine it's not, but maybe it is.

I would imagine you could probably record infinite values, though.

Um, but I really don't know the answer to that.

Um, sorry for the live test, but I'm actually kind of curious because it's fairly useful if you can record a bunch of things at once.

I'm pretty sure the limit on quant O P in is ah.

I just want to confirm that I'm sorry if you guys aren't really interested.

I assume if you're watching this, you probably are at least somewhat interested in this recording.

I just wanted just curious That followed through.

So, anyone around those again before we're dealing with those?

So these are all the other columns that you have.

If you wanted to when you were going, you could do stuff you could.

Also, just after the fact that you could make, um you could make calculations on this data frame, like you can define a new column, and you could see that new columns equal the benchmark period return minus algorithm, period return or something like that.

You can always do stuff like that are really proud to algorithm, period minus benchmark period.

So, like, that would be, um you could make those kind of like after the fact calculations, if you don't know anything about pandas.

戦略メトリクスの可視化 - Ziplineチュートリアル Pythonによるローカルバックテストとファイナンス p.2 (Visualizing Strategy Metrics - Zipline Tutorial local backtesting and finance with Python p.2)