Building a Trading Robot in Python | Pt. 1

Summary

In this video, the creator introduces a series on building a trading robot using Python, specifically with the TD Ameritrade API. Aimed at viewers with a TD Ameritrade account and a solid understanding of Python, the series will guide through the process from start to finish. The script emphasizes the importance of understanding the underlying mechanics of the robot, including handling price data, managing account activity, and calculating technical indicators. The presenter outlines the need for objects like the robot, stock frame, portfolio, trade, and indicator objects, and mentions that the library is available on GitHub for cloning, with future episodes promising a detailed, hands-on coding approach.

Takeaways

• 😀 The video introduces a series on building a trading robot using Python, specifically with the TD Ameritrade API.
• 🔑 It's assumed that viewers have a TD Ameritrade account, but the presenter may cover other APIs in the future.
• 🛠️ The tutorial is aimed at individuals with a solid understanding of Python and familiarity with libraries like pandas and numpy.
• 🚫 The presenter advises beginners or those not comfortable with programming to hold off on this series due to the complexity.
• 💡 The series emphasizes the importance of understanding the underlying mechanisms of the trading robot for effective use.
• 📈 The robot will need to fetch price quotes, manage account data, enter and exit positions, and calculate technical indicators.
• 🧩 The script outlines the creation of several objects: Robot, StockFrame, Portfolio, Trade, and Indicators, each with specific roles in the trading process.
• 💼 The 'Robot' object acts as the primary interface with the TD Ameritrade API, handling tasks like getting quotes and placing orders.
• 📊 The 'StockFrame' object is designed to organize and manipulate price data, making it easier to work with financial data in pandas.
• 💰 The 'Portfolio' object mimics account activity to aid in decision-making, such as avoiding duplicate positions.
• 📉 The 'Indicators' object is intended to simplify the calculation of technical indicators, ensuring consistency and flexibility.

Q & A

What is the main focus of the video series being introduced?

The main focus of the video series is to guide viewers through the process of building a trading robot from start to finish using Python, specifically with the TD Ameritrade API.

What is a prerequisite for viewers before starting the series?

Viewers should have a TD Ameritrade account, and it is assumed that they have a relatively decent understanding of Python and familiarity with libraries like pandas or numpy.

Why is it important for viewers to understand what's happening behind the scenes when building a trading robot?

Understanding what's happening behind the scenes helps viewers grasp how the robot is structured, how different objects interact, and why it was built in a certain way, which is crucial for effectively using and potentially modifying the robot.

What are some of the complexities involved in trading that the video series aims to address?

The series addresses complexities such as understanding order types, technical indicators, and account activity, all of which are essential for effective trading and are incorporated into the programming of the trading robot.

How does the presenter plan to make the process of building a trading robot more approachable?

The presenter plans to make the process more approachable by starting at a high level, defining objectives, and then gradually diving into the programming aspects, ensuring that viewers understand the purpose and functionality of each component.

What is the significance of the 'robot' object in the context of the trading robot being built?

The 'robot' object is significant as it mimics the user's interaction with the TD Ameritrade account, handling all communications with the TD Ameritrade API, including getting quotes, placing orders, and retrieving account activity.

Why is the 'stock frame' object necessary in the trading robot?

The 'stock frame' object is necessary for organizing price data in a consistent and finance-oriented manner, making it easier to add, delete, select, and manipulate data for trading decisions.

What role does the 'portfolio' object play in the trading robot?

The 'portfolio' object mimics the user's account activity, helping to manage and organize information about current positions and active orders, which is crucial for making informed trading decisions.

How does the 'trade' object simplify the process of managing orders in the trading robot?

The 'trade' object simplifies order management by providing methods to easily modify orders, add components like stop losses, and check the status of orders, making the process less complex and more streamlined.

What is the purpose of the 'indicator' object in the trading robot?

The 'indicator' object is designed to calculate and define technical indicators in a consistent and easy-to-use manner, ensuring that the robot can use these indicators effectively for trading decisions.

Where can viewers find the current state of the trading robot library discussed in the video?

Viewers can find the trading robot library on the presenter's GitHub account under the repository named 'Python trading robot', where they can clone the repository and explore the code.

Outlines

00:00

🤖 Introduction to Building a Trading Robot with Python

The speaker introduces a new series focused on building a trading robot using Python, specifically with the TD Ameritrade API. It's mentioned that the series has been highly requested and will be structured to go from start to finish in creating the robot. The speaker emphasizes the importance of understanding the underlying processes of the robot, not just using the library. It's assumed that the viewer has a TD Ameritrade account and a decent understanding of Python and libraries like pandas or numpy. The speaker also advises those new to programming to familiarize themselves with basic programming concepts before proceeding with the series.

05:01

📈 The Importance of Defining Trading Objectives and Details

The speaker discusses the importance of being explicit about trading objectives and the details of trading strategies when building a trading robot. It's highlighted that many people struggle with defining their trading methods precisely, which becomes necessary when programming a robot. The speaker suggests that traders should write out their trading strategies to understand the finer details, such as order types and quantities. The paragraph also touches on the need to understand the big questions and objectives of the trading robot, such as obtaining price information, entering and exiting positions, and managing account data.

10:02

💼 Necessity of Account Activity and Technical Indicators in Trading

The speaker continues by emphasizing the need for a mechanism to obtain account activity and the importance of calculating technical indicators for trading decisions. It's explained that account activity can influence trading logic, such as not entering a new position if one already exists for the same ticker symbol. The speaker also discusses the challenges of calculating technical indicators and the intention to create a consistent and flexible way to calculate them through the use of an indicator object. The paragraph outlines the high-level objectives for the trading robot, focusing on the need for price information, position management, account data handling, and indicator calculations.

15:04

🛠️ Designing the Robot's Objects for Effective Trading

The speaker delves into the object-oriented programming approach for the trading robot, starting with the robot object that mimics the user's interaction with the TD Ameritrade API. The robot object is essential as it handles all interactions with the API, including getting quotes, placing orders, and retrieving account activity. The paragraph also introduces the concept of a stock frame object to organize price data and a portfolio object to mimic account activity. The speaker discusses the benefits of having these objects, such as ease of data manipulation and decision-making based on account data.

20:04

📊 Streamlining the Trading Process with Trade and Indicator Objects

The speaker further elaborates on the design of the trading robot by introducing the trade object, which encapsulates the complexity of orders, and the indicator object, which simplifies the calculation of technical indicators. The trade object is designed to handle modifications and additions to orders easily, while the indicator object ensures consistent and flexible calculation of indicators. The speaker also discusses the importance of having the stock frame and indicator objects communicate effectively, with the indicators recalculating as new data is added to the stock frame.

25:07

🔗 Accessing the Trading Robot Library and Preparing for the Next Steps

The speaker wraps up the presentation by directing viewers to the GitHub repository where the trading robot library is available for cloning. It's mentioned that the library is a work in progress and will undergo significant changes. The speaker provides instructions on how to clone the repository and gives a brief overview of its contents, including the main robot file, configuration requirements, and examples of how to use the library. The speaker encourages viewers to contribute to the project and hints at the next video, where the creation of the robot will begin from scratch.

Keywords

💡Trading Robot

A trading robot refers to an automated system or software that executes trades on behalf of a user based on predefined parameters and strategies. In the context of the video, the main theme revolves around building a trading robot using Python, specifically for the TD Ameritrade API. The video aims to guide viewers through the process of creating such a robot from scratch, highlighting its importance in automating and executing trades efficiently.

💡Python

Python is a high-level, versatile programming language that is widely used in various fields, including data analysis, artificial intelligence, and web development. In the video, Python is the chosen language for building the trading robot due to its readability, extensive library support, and its suitability for financial applications. The script mentions that viewers should have a decent understanding of Python to follow along with the coding process.

💡TD Ameritrade API

The TD Ameritrade API is a set of programming tools provided by TD Ameritrade that allows developers to access and interact with the brokerage's trading platform programmatically. The video series focuses on utilizing this API to build a trading robot that can fetch real-time market data, place orders, and manage account activities. The API is central to the robot's ability to perform trades and manage account information.

💡Pandas

Pandas is a powerful Python library used for data manipulation and analysis. It provides data structures and functions needed to work with structured data. In the video, the speaker mentions that the trading robot will use Pandas for organizing and manipulating financial data, such as stock prices and account activities, which are critical for making informed trading decisions.

💡Technical Indicators

Technical indicators are statistical tools used by traders to analyze and predict the price movements of financial markets. They are based on an asset's historical price, volume, or open interest. In the video, the speaker discusses the importance of calculating technical indicators as part of the trading robot's functionality, as they are used to generate trade signals and assess market conditions.

💡Object-Oriented Programming (OOP)

Object-Oriented Programming is a programming paradigm that uses objects and classes to design applications and complex software systems. The video script mentions that the trading robot will be built using OOP principles, which allow for a modular and organized approach to programming. This includes defining objects such as 'Robot', 'StockFrame', 'Portfolio', and 'Trade' to encapsulate specific functionalities and data.

💡Order Types

Order types refer to the various ways in which a trade can be executed in the financial markets. These include market orders, limit orders, stop orders, and more. The video script emphasizes the need for the trading robot to understand and execute different order types as part of its trading strategy, which is crucial for managing risk and ensuring that trades are executed as intended.

💡Account Activity

Account activity encompasses all the transactions and positions within a trading account. In the context of the video, managing account activity is a key function of the trading robot, which needs to track current positions, open orders, and other relevant data to make informed trading decisions and avoid conflicts, such as entering multiple positions in the same security.

💡GitHub

GitHub is a web-based platform for version control and collaboration that allows developers to work on projects, manage code, and collaborate with others. The video script mentions that the source code for the trading robot will be available on GitHub, allowing viewers to clone the repository, review the code, and even contribute to its development by submitting pull requests.

💡Unit Testing

Unit testing is a software development method where individual components of a program are tested to determine if they correctly perform their intended function. In the video, the speaker mentions that there is some unit testing included in the robot's codebase, which is important for ensuring that each part of the robot works as expected and for identifying and fixing bugs during development.

Highlights

Introduction to a series on building a trading robot using Python and the TD Ameritrade API.

Assumption that viewers have a TD Ameritrade account and a basic understanding of Python and libraries like pandas or numpy.

Emphasis on the importance of understanding the underlying processes of the trading robot for effective usage.

Advice for beginners to gain more programming knowledge before engaging with the series.

Explanation of the complexity involved in trading and the additional layer of complexity introduced by programming.

The necessity to define clear objectives for the trading robot at a high level.

The importance of being explicit about trading strategies and order types when programming a trading robot.

The need for a mechanism within the robot to obtain price information from the market.

Discussion on the functionality required for the robot to enter and exit trading positions based on signals.

The role of the robot in managing account data and its influence on trading decisions.

The significance of calculating technical indicators and their impact on trading decisions.

Introduction to the 'robot' object, which mimics user interaction with the TD Ameritrade account.

Explanation of the 'stock frame' object for organizing and managing price data.

Description of the 'portfolio' object that reflects account activity and assists in decision-making.

The 'trade' object facilitates complex order management and real-time trading activities.

The 'indicator' object standardizes the calculation of technical indicators for consistency.

Availability of the library on GitHub for cloning and potential contributions.

Instructions on how to clone the repository and the current state of the library with its dependencies.

Overview of the main robot file, configuration requirements, and examples of how to use the library.

The intention to rewrite the library from scratch in the next video, showcasing the setup and core functionalities.

Transcripts

well welcome back so I know a lot of

people will be excited about this series

we are kicking it off yes we are gonna

cover a series that has been highly

requested by so many people we are gonna

be showing you how to from start to

finish

build a trading robot using Python and

more specifically we're gonna be showing

it using the TD Ameritrade API so yes

this again just putting that out there

right now this will be assuming that you

have a TD Ameritrade account that is not

to say down the road I will not do

Interactive Brokers or the other ones

but right now this one week covering

purely TD Ameritrade so this wonderful

series it's been a few months in the

making so there's a lot of stuff that

kind of comes along but this I will say

that additionally I do want to spend a

few minutes just going over it right so

I don't want to jump right into the

coding I really want to take some time

let's talk about what are we trying to

do and how are we trying to do it for

those of you who I know there's a lot of

people out there they kind of just want

to use the library they don't want to

necessarily kind of care about what's

going on behind the scenes that is okay

to a certain extent but I will say this

you really should know kind of what's

going on behind the scenes the only

reason I say that is you know there's a

lot of things that are being calculated

and it's a little bit easier to

understand how to use it if you kind of

understand how it was built so even

though things will probably change a

little bit from what you're gonna be

seeing now it's gonna change a lot but I

think having that idea of like hey how

did we structure the code how do we

think about the different objects and

how do they kind of all interact

together yeah you might not necessarily

go out there and rewrite the library but

it's a good idea to kind of just

understand why was it built this way and

kind of how's everything talking to each

other additionally I will say this is

well it is assumed that if you're

watching this series

that you're coming in with you know a

relatively decent understanding of

Python and some level of familiarity

with certain popular libraries like

pandas or numpy you know something like

that if you're very new to programming

I'm actually gonna say hold off on this

series for a little bit you know you

might want to go get familiar with

certain concepts like you should be

comfortable with you know class objects

you should be very comfortable with if

statements for loops anything like that

I know sometimes we kind of just want to

jump into it but the thing is this can

actually make real trades for you and if

you don't understand how it's working

and if you don't know how to structure

the code you're gonna be making real

trades with real money so with that

being said you know I am saying if

you're not comfortable with programming

in that regard you need to get familiar

with it and then kind of come back to

this series I do have some videos that

cover different topics but there are

also plenty of other channels out there

who have a ton of content that will more

than likely get you right up to speed so

with that being said let's get started

all right first things first this is

actually a really complex topic trading

is already hard enough right like a lot

of people who day trade who just even

you don't even really day trade I mean

just trade in general you know you learn

right away that it's not easy right I

mean there's a lot of things you have to

understand there's all these fancy terms

there's all these technical indicators

out there and there's all this kind of

background knowledge that you have to

understand in order to trade effectively

so you've got to understand order types

you've got to understand indicators if

we got to understand you know different

things related to account activity so

we've already taken a pretty complex

topic and now we've made it even more

complex by adding this whole wonderful

world of programming into it so that

just makes a little bit more challenging

doesn't make an impossible and doesn't

even make it where it can't be done it

just means we need to be very focused

let me go about this whole process and

part of being focused means we really

need to start at a high level by

defining our objectives

what I have found is what I found is a

lot of times people will reach out and

say hey I'm trying to build a training

robot and I've kind of picked their

brain a little bit and I say okay well

what are you trying to do like how are

you trying to do it and something that

kind of became very apparent to me

relatively quickly was people were

trading but they also weren't

understanding how they were trading in

very explicit terms because behind the

scenes they were making all these

assumptions right and certain things

like you know when they would make an

order you know for a lot of people maybe

for example a big assumption was it's

just a market order well now all this

sudden I'm gonna have to you're gonna

have to ask I'm gonna have to ask you

that right I'm gonna have to you're

gonna have to tell me what type of order

you want to place so all the sudden I'm

gonna start requiring to be very

explicit with how you trade because you

don't get that luxury anymore

you have to be able to tell that robot

hey you need to trade like this and if

you want it to truly mimic your trading

you have to give it all that information

and you have to tell it upfront so hey

you know it got to be this order you've

got to be using this price you've got to

be you know buying this amount of

quantity so all of the sudden this is

where people tend to struggle a little

bit because now I'm forcing them to

really be extremely explicit about how

they're trading so a lot of times what I

have found is having people take a step

back and really sit down right I mean

write it out I mean there's nothing

wrong with writing out but write out how

do you trade you know it's not just

sitting there well I just placed an

order it's more than that and you know

that it's you know there's a type of

order there's a quantity you know there

there's all these little finer details

that are going into it and you now need

to start identifying those finer details

so that's usually where again for at

least from my experience this is where

people kind of struggle a little bit

because hey I'm forcing them be very

explicit and it's very easy to get

overwhelmed so naturally what I do is I

kind of take a step back and say okay

well you don't have to answer every

question but ideally you want to have a

general idea about kind of the kind of

the big the big questions right

the same kind of relates here when we're

talking about the objective right we're

not gonna have to necessarily define

every little fine detail yet but we

should generally have a decent

understanding about what are the big

things that this trading robot needs to

do now this is not an exhaustive list

this doesn't contain every possible

thing it needs to do but when I was kind

of approaching this I kind of asked

myself what is really the things that

this thing needs to do right so what are

the objectives of this training robot

and what does it have to do in order to

function the way I think it should

function well the probably the biggest

thing is it needs to string quotes right

so it needs to be able to get price

information from somewhere because that

price information is used in so many

different parts it's used when we're

calculating indicators it's used when

replacing orders it's used when we're

trying to organize our data so one of

the first things we have to have built

into this is we've got to have some type

of mechanism for getting price

information once we have that price

information we can manipulate it and do

whatever we want with it but we have to

have some type of mechanism for getting

that information now right now the way

I've built the library is I am using not

this TD streaming API I'm using the TD

standard API so it's not to say the

streaming API won't come in at some

point but right now just to kind of keep

things a little bit more simple I'm

relying purely on the standard API and I

will show you there's ways we can kind

of mimic that's a loose word but we'll

stick with it mimic the streaming API in

certain regards so the big first thing

we need to have is price information so

we're gonna have to basically have a

robot somehow get price additionally we

need some type of functionality where we

can enter and exit positions right

that's trading you're gonna enter a

position because you got a buy signal

maybe and then maybe you're gonna have

to exit a position because you got a

sell you know a sell signal right so we

have to have some type of functionality

those robots got to be able to hey take

a signal and then you know act on that

signal and the way it acts is it

determines well maybe I need to create a

new

order that will then reverse the

position I just created right so we have

to have some type of functionality like

that additionally we need to have some

type of functionality where it goes in

it can help us manage our account data

right and so we have a lot of account

data out there and very often we will

find that that account data is critical

when it comes to making trades for

example I have some individuals who will

say once I own a position I do not want

to enter another position for that same

ticker symbol so we now have to build in

some logic that says hey the minute that

person enters a position even if I get a

buy signal do not enter a new position

so this is all of the sudden we're being

explicit is very important because

sometimes people don't realize that I

some people won't trade like that and

some people say I don't care if there's

another bias appeal maybe we just add to

the position not saying it's right or

wrong but I'm saying as you have to

define that role beforehand so because

of that and because sometimes the logic

that we build depends on what we already

have in our account activity we need to

have some type of mechanism for getting

that account activity and organizing it

in a fashion where we can then use it in

other decision-making processes of our

robot and probably the big one is we

need to have a mechanism for calculating

technical indicators a lot of

individuals use technical indicators in

order to trade they use it to determine

when to enter or exit a position they

use it to determine when the market is

behaving more in a more active manner

compared to a more you know I don't want

to say it's a sedative is that what a

sedative mean I don't know regardless

they use it in order to make decisions

sometimes it's determined when to start

looking for signals other times it's use

to determine when to enter or exit a

position so we have to have some type of

mechanism for helping us calculate those

indicators so with that we have an

objective

this is ideally at a very high level

this is what we want our robot to do

when we think about our solution we now

have to kind of start going into our

programming world right well a lot of

times we're thinking a very

object-oriented programming so with that

being said we need to start thinking

about our objects what are the objects

that we're gonna be creating and what

are they gonna be doing for us like how

does it fit into the bigger picture of

our robot well the first thing that

we're gonna have to do is we're gonna

have to mimic guess what you write so we

have to mimic you as an individual the

way we mimic you is we define our robot

object our robot object is mimicking how

you interact with your TD Ameritrade

account but what this robots gonna do is

it's gonna interact with the TD

Ameritrade API so the robot is really

designed to mimic you and the way it's

gonna be mimicking you is it's gonna be

making requests to the TD Ameritrade API

right so it's gonna be getting quotes

it's gonna be placing orders it's gonna

be getting account activity so this

robot is gonna be handling all the

talking to the TD Ameritrade API and so

the way I also like to think about this

is this is the highest level of our kind

of hierarchy if we think about our

objects as a hierarchy we can't do

anything unless we have a robot so we

can't do any trading unless we have you

in the picture so even though you're not

necessarily there physically

manipulating the robot the robot is

designed to mimic you so it's like we're

copying you or putting you into a

program and we're saying hey this robot

has to be the main point of entry if I

don't have a robot I can't do anything

else so that's our first object our

second object we're gonna be working

with a ton of price data and we are

going to need to organize that price

data in a very consistent manner and

we're going to have to have easi

mechanisms for adding data appending

data well I guess

adding an appending are the same thing

maybe in some cases deleting data

selecting data and just keeping all of

our information organized in a very easy

fashion so we're gonna create a object

but we're gonna call a stock frame now

maybe a stock frame isn't the best name

and I've actually debated if that's even

remotely correct but for right now we're

gonna keep it like that the stock frame

at the end of the day is gonna be just a

panda's data frame but I don't want to

make it too general right I don't want

to necessarily say oh yeah it's a

panda's data frame so why are you doing

all this extra fluff and I debated about

that I mean initially I said you know as

am I just doing overkill right now like

is it really necessary to just define a

new object that really is just a panda's

data frame and I found it kind of is

sometimes in the only reason why it is

our data is going to be very finance

oriented right and there's a lot of

finance operations that you can do in

pandas that would just be nice or a

little bit easier to do if it was set up

correctly right so things like you know

maybe it makes sense to have a multi

index data frame instead of just a

normal data frame where you know the

index is all right there so that might

necessarily be beneficial

additionally there are some mechanisms

where adding data to a data frame could

be a little bit challenging sometimes

depending on you know how its organized

and so I wanted to kind of take away the

hassle of managing that and just say you

know what as long as we're interacting

with the API consistently all I really

care about is doing common operations so

maybe that's adding new rows right and

then once I've added new rows there

might be certain other things I might

want to do right so for example if I

have indicators in my stock data frame I

might want to refresh those indicators

or in other words recalculate them with

the new data in mind and so what we'll

find is that stock frame is going to

make that process very easy for us very

very easy for us so you don't have to

worry about that now you're just gonna

call a method and guess what it's gonna

do a lot of magic behind the scenes but

that's what it

to do it's supposed to make it easier to

work with so the stock frame is gonna

really be helping us to organize our

data additionally we're gonna have our

next object which is a portfolio our

portfolio is designed to kind of mimic

our account activity

maybe that's the orders that we already

have active right or maybe that's the

positions that we have currently in our

account that are have been filled right

so the portfolio is really designed to

mimic our account activity because a lot

of times we'll be making decisions based

on data that comes from our account

right so say for example I don't want to

enter a position if I already have that

stock already in my portfolio well this

way by having a portfolio object we can

easily ask those questions and then

transfer that information to our

decision making process so with our

portfolio object it will be very useful

when it comes to organizing that

information and asking questions as

we're trading that's really how I think

about is that asking questions component

do I have that stock well let me just go

check my portfolio and then if I do oh

great I'll just go trade I don't okay

well no I won't trade so that's really

important additionally there's some

times where our trading can really

depend on certain metrics about report

folio right so for example maybe I don't

want more than 60% of my portfolio in

equity right so it would be nice if we

had some type of free built

functionality where we could easily

calculate those metrics on the fly and

so ideally with this portfolio object we

would be able to answer those questions

and then calculate those metrics as we

need and then additionally as we may

might need to do more advanced

operations we can now easily add that

functionality but keep it contained to

where it belongs and that really is that

portfolio aspect so that's the portfolio

object next object trading right so

we're gonna be doing a lot of training

so we need to define a trade object

right so our trade is really our order

but we also understand that an order can

be relatively complex depending on how

you're trading it might not just be a

market order it could be a very it could

be no

you order it could be a child order

incorporated into it so all of a sudden

we have to have some type of structure

where we can easily modify our orders or

we can easily add components to it in an

easy fashion and this is normally where

people struggle a lot because

unfortunately the TD Ameritrade API was

really lacking when it came to the

examples of how to place orders and how

to do that and so over the you know many

months of using this API I've kind of

figured out ok well this is how you do

that and I wanted to kind of have that

functionality pre-built for you so that

way if you want to add things like a

stop loss or a stop loss and take profit

just just an order you already placed

well you can just call a method and it

will do all that for you behind the

scenes instead of you having to take

these very nested dictionaries that can

get very messy very quickly and instead

use that trade object to do it and then

additionally what I would like to do is

I'd like to see that trade object help

us answer questions like has it been

filled yet has it was it partially

filled so ideally I'd like this trade

object to help us answer even more

complex questions that might come down

the road that functionality might not be

there right now but I could definitely

envision that object actually containing

it

so that's why I thought it was important

to help encapsulate all that into that

kind of just that single object because

it's very important and it really is

kind of I would say the meat of our kind

of real-time activity we want to be able

to ask those questions on the fly and so

that trade object will help us do that

final object indicators indicators we've

already said this multiple times a lot

of people use technical indicators in

order to make trades now right now we

are purely doing it from a technical

analysis standpoint that is not to say

that down the road we might not include

include a fundamental analysis but right

now it is purely technical analysis and

so this indicator object is going to

help us calculate and define

those indicators for us in an easy

consistent fashion I think the biggest

thing here is the consistent fashion

it's very easy to get overwhelmed when

it comes to calculating indicators and

it's very easy to find yourself in a

position where all the sudden you're

working with this pandas dataframe and

it gets really hard to do it because

what was it grouped was it not grouped

was it organized was it not organized

you know and all of a sudden you will

start pulling your hair out just like I

did because all of the sudden it's not

straight forward and that was like the

thing that drove me crazy so what I

found was let's define an indicator

object where we can easily calculate

those indicators and as long as I know

what you're passing me through hint hint

it's the stock frame but as long as I

know what you're passing through

I can easily calculate all those

indicators for you in fact initially

what I was actually gonna do is I was

gonna use a few libraries to to

calculate those indicators cuz I thought

man this is great it's like that's gonna

be super easy no problem whatsoever and

then what I found out was it was always

assumed that you were giving a very

standard data frame and I said well

that's assuming the columns are named

the same that's assumed that this is

named the same and all this other stuff

and that it's organized and then they're

assuming that when it's going in there

that it's sorted and well I said well

what if it's not sorted are they gonna

sort that for you and so all the sudden

I found man this could actually be a

little bit challenging because I'm gonna

have to do that on my end because I

don't know if they're necessarily doing

that on their end and so it's not to say

maybe down the road I can't integrate

those libraries but right now I had to

actually define some of these indicators

myself because I was finding that I know

they weren't nest they were assuming

certain things when it was being put in

there and so for example that I think

the biggest thing was how do you

calculate it if you have multiple what

is it multiple ticker symbols in the

same data frame a lot of those libraries

can't handle that and so all this I said

man if I pass through groups it's not

going to work I said I'm gonna have to

filter or do XYZ and I said man this

this could get challenging really

and so the indicator client object was

really designed to kind of help make

that interaction consistent but still

give you the flexibility where you can

have all your data in a single data

frame additionally a lot of indicators

can be calculated using different

methodologies I want that flexibility

for the user so I want the user to be

able to say hey I want to calculate the

RSI using this methodology great just

tell me you want to do that and then

once I have that information

I can calculate that for you so that was

really the intent of the indicator

object was make it consistent and make

it ideally easier to calculate those

indicators real time and then probably

the biggest thing was having that stock

frame and those indicator objects talk

to each other so as the stock frame was

adding new data the indicators would

recalculate that is very challenging to

do sometimes in pandas and honestly that

was one of the one things that I think

always surprised me about pandas and

maybe it could just be the fact that I

just maybe I'm seeing the answer right

in front of me and I'm just not seeing

it but something where it's like I'm

just adding rows like how are these

calculated columns not updating and so I

basically created what I'm calling is a

refresh method and basically what I've

done is when you call your indicator

client I've saved all the arguments

you've passed through so things like the

period that you use the methodology that

you specified and I save that in the

indicator client itself and so what

you'll find is when I call the refresh

method I'm literally calculating those

indicators again but I'm using the saved

information you sent me when you

initially initialized those indicators

so it's pretty neat honestly I out I

thought that was kind of you know that's

this is maybe where Python really shines

a little bit but I thought that was kind

of cool so at a high level this is how

we're gonna start thinking about

building our solution so with that being

said though that actually finishes our

presentation only three slides but of

course I was rambling no there's that

too

so for those of you who are interested

in it on the library is currently posted

on github

so you are more than welcome to clone it

and is not on pi PI so you cannot

install it at least not yet it does have

a few dependencies so pandas numpy and

TD Ameritrade big one is pandas I'm

pretty sure once you install pandas you

don't technically need this but I'm just

thinking out there just to be safe but

really if you want to kind of just

high-level like what does it look like

well first thing is you want to go to my

github account so a breed 1 1 9 2 then

you can see this wonderful photo which

is horrible and then you will see some

repos that I currently have available

and one of them is called the Python

trading robot and so this is the one

that I will be recreating in a series

again just to be very front lots of

changes coming so this will change a lot

so don't be surprised if you go in here

ever you're so off you're like wait he

changed it again oh yeah lots coming and

then from here if you want you can

always clone it so if you just go over

here to the green little button clone or

download and then if you copy it you can

just do it's like so I've covered that

in other videos so I will put a link to

one where I've showed how to clone a

repo but again hopefully pretty

straightforward there's some unit

testing kind of built in I'm adding to

it but right now there's some stuff in

there most of it is for the main robot

itself so just you know again asking

some general questions and then there's

some indicators portfolios and you know

stuff like that and then documents well

unfortunately nothing right now but at

least a template and then you can

actually see the main robot file so if

you were running it this stuff again I

do have a config file I get this asked a

lot a lot a lot a lot you need you don't

have to do this part you don't have to

but you need to have your client ID and

you need to have your redirect URI and

you need to have your credentials path I

am NOT you have to provide that

information for you how you choose to

provide it is totally up to you you can

hard-coded in your script I don't

recommend that if you're sharing it with

other people um you really should keep

it secure

but you can't so I have a config file so

I write it every time I create a new

library and then I just import it as I

need it and then you'll see here I

create a robot and then I do things like

add positions just ask certain questions

and stuff like that so ideally this

should be a little bit easy to hopefully

follow I would hope but sometimes not

and then um again you can go through

here I've been trying to document

everything so I've been trying to add

doc strings and then I'm gonna try to

use like usage cases as well so I'm

trying to make sure that's all organized

if I've missed one you're more than

welcome if you want to either bring it

up or just do a pull request and then

just add it for me but yeah so I mean

there's lots of good stuff here

hopefully again I'm trying to be as

clear as possible about how to use it

so hopefully it's it's hopefully

straightforward that's that's the goal

at least but I'll be probably again just

making lots of changes so again please

don't be surprised if you come here when

you're like geez he just changed like

five files but yeah so I mean you can

obviously clone it by all means if you

have suggestions very open to it people

have been contributing to the

Interactive Brokers one the TD

Ameritrade one so again people have been

great with that and so I'm always open

to hearing that might not get back right

away but always hoping to hearing some

suggestions other than that I will say

I'm gonna think that's gonna be it for

this particular video next video I will

be writing everything from scratch so I

will not be cloning it we will just be

basically recreating what we're seeing

here - the setup and a lot of this other

fancy stuff but for the most part we

will be creating everything that we're

kind of seeing right here so that's

what's gonna be coming up in the next

video you got any questions at this

point by all means ask away hopefully I

have an answer if I don't I'm sorry but

yeah that's pretty much it so hopefully

everyone's happy now we're doing a

trading robot

that's the exciting part right and then

hopefully hopefully it's profitable

right

we'll see but yeah all right we'll see

you in the next video