Algorithmic Trading Using Python - Full Course
HTML-код
- Опубликовано: 2 июн 2024
- Learn how to perform algorithmic trading using Python in this complete course. Algorithmic trading means using computers to make investment decisions. Computer algorithms can make trades at a speed and frequency that is not possible by a human.
After learning the basics of algorithmic trading, you will learn how to build three algorithmic trading projects.
💻 Code: github.com/nickmccullum/algor...
✏️ Course developed by Nick McCullum. Learn more about Nick here: nickmccullum.com/
⭐️ Course Contents ⭐️
⌨️ (0:00:00) Algorithmic Trading Fundamentals & API Basics
⌨️ (0:17:20) Building An Equal-Weight S&P 500 Index Fund
⌨️ (1:38:44) Building A Quantitative Momentum Investing Strategy
⌨️ (2:54:02) Building A Quantitative Value Investing Strategy
Note that this course is meant for educational purposes only. The data and information presented in this video is not investment advice. One benefit of this course is that you get access to unlimited scrambled test data (rather than live production data), so that you can experiment as much as you want without risking any money or paying any fees.
This course is original content created by freeCodeCamp. This content was created using data and a grant provided by IEX Cloud. You can learn more about IEX Cloud here: iexcloud.io/
Any opinions or assertions contained herein do not represent the opinions or beliefs of IEX Cloud, its third-party data providers, or any of its affiliates or employees.
I like how the errors were made without editing them out from the clip. Thanks for including all those in.
I LOVE how you go through your own process of debugging. It really helps me.
I don't comment often but this guy is an amazing teacher. He explained everything in such good detail. It helped me as I need a deep understanding of a topic to grasp it completely. Keep up the great work.
Not only application specific, he goes through very comprehensive instruction on Python. Huge kudo!
Still watching and learning the topics, so far so good, thanks a lot for generous offering this great course and I look forward to seeing this extended topic with advanced techniques and link to specific brokers' API like IB, Futu, etc. Thanks again for your great teaching!!
Excellent course, NIck!! I enjoyed very much the "real life" approach that included the footage of tracking down bugs. Thanks for not editing that out!
50min in and am already in love with the style of teaching. Errors are left in the vid as well as the logical process of debugging them, everything he mentions are backed up by examples. Superb content
where did he has shown how to install dependencies and cloning plz help
Thank you that you didn't edit video, these debugging/looking for error processes actually make things more clear. And teach how to deal with such cases.
And thank you for this video at all, it's great!
Excellent course for someone new to Algo Trading. It’s precise and well designed course. I would like to see something more along lines of simulating real time strategies like VWAP, TWAP. This one is more of pre-trade analytics. Nonetheless I liked it and strongly suggest to anyone new to Algo trading space!
Oh my god, again, you are doing god work. From dynamic programming now this.
this was awesome and though I've just finished the first part only, I cannot thank you enough for how well you explained each portion of the project. Now only, if I could build up from here!
Great course. Thank you for all the work you put in and your thorough explanation of how it all works together while coding. I haven't been coding python for more than a year and this was a great refresher course into data science and api use
where did he has shown how to install dependencies and cloning plz help
thank you so much for this course!!! I started my coding journey with Free Code Camp, I now work full time as a developer. This topic is on my to do list!!!
Which road map did you take for fulltime job?
Cap af
great video, what I really liked is that you showed in a super clear way how to use the requests library in Python and now I do not need to depend on 3rd party wrappers that might not have all functionality! thanks!
This is a great course. I learned so much from it. One suggestion: the robust value trading program should exclude the negative PE Ratio stocks first before sorting for the low 50 stocks.
Thanks Nick for the complete and detailed course for beginner. Really appreciate.
Sir, you clearly explained both financial and coding keywords better than my professors did in 4 years of schooling. You're a great teacher!!
So ageee!!!!
This is the most common comment you could ever read on web courses. I’m sure your teacher explained well, you were just bored and didn’t pay attention. Here you have 4 hours for everything you need to know and you can stop and start whenever you want. It’s easier. But don’t discredit other people because you have not been able to keep up with them.
@@AlbyTheMovieCreator the school system is just fcked and you know it
I wonder why almost everyone feels the same about the schools.they kepp u 4 years in a box and u act like a piggy bank to them.it is realy sad .I am already in that box.this guy deserves a ton of money not my teachers.
Everything covered in this video is extremely basic. If you didn't understand it in 4 years of college, that is 100% on you.
Congrats Nick McCullum for creating this. Good that the debugging parts are left in, so we can understand in more details how various parts work.
Thank you for this amazing tutorial! I wish there was more of this information 4-5 years ago.
Thank you very much for this course!!! You've done a wonderful job!
freakin hell! how good is this! Lockdown boredom solved. Thanks!
finished the entire thing in one sitting. Loved it! its an amazing course if you are starting out algo trading!
Worth hai kya?
In 58:05 for anyone don't want to use yield, you can use list concatenation for both symbol_groups and symbol_strings
symbol_groups = [list(stocks.Ticker)[x:x+100] for x in range(0, len(list(stocks.Ticker)), 100)]
symbol_strings = [(',').join(i) for i in symbol_groups]
do u have any tips on how to learn and improve my ability in this area after I complete this video? Tryna prepare myself for work in this field was wondering if you have tips!
So far I have only used my python knowledge to program basic games, automation widgets and a few data scrappers but this... My head just exploded with curiosity!🤯 Thank you so much!
@Arid Sohan *paid
@@Mmmkay.. I don’t think that was a typo
@Arid Sohan tf
ya maybe program to teach a bot to give head , we can control the speed
29:41 Thanks for showing debugging when hitting errors. That adds a lot to real world training!
that is not debugging lmao, it's like turning your phone off and on
Reloadi g this thing is like off and on but how do you know if you should restart it? This is debugging
Its still showing error importing from secrets.py file. Did nothing for me
hey guys, even after restarting kernel, I am still getting same error for importing secrets, any solution?
@@koushikdey7778 same here please do suggest
I'm 12 minutes into this course and I already feel like I'm getting a firm grasp of this stuff. Thank you so much
Stop picking low hanging apples. Go deeper into the subject.
@@panhuragan4388 😂😂
This is very great video filled with explanation and experience. Watching it has been a pleasure. Thank you for taking the time and having a student mentality you are appreciated.
I like how you face problems and fix them, thanks for not cutting those parts out of the final video
to whom it can be of use: if you're stuck because of a KeyError 'DISCA', that is because the original stock list has had some change. You can solve it by using a try/except block when doing the the iteration for the final_dataframe = final_datafrme.append(.....) . Just put all of that inside the try block, and then do an except KeyError, and handle the eroor
haven't used a try/except before, can you please elaborate on how to use the except? I have the same issue
Edit: I just put everything in the try block like you said and put except KeyError: break, is this what you did?
Thank you for this video but I think everybody is feeling the absence of the second section. Any updates would be greatly appreciated.
This guy is awesome, the best thing I liked that he was debugging things in video. Debugging is actually most important part that takes time to learn.
@2:09:00 you can use the code below to ensure you have a valid number by looping until the user has inputted a numerical value. Thanks for the great video!
while True:
try:
portfolio_size = float(input("Enter the size of your portfolio: $"))
break
except ValueError:
print("Please enter a valid numerical value in $")
Hey thanks bro you saved the day :)
Bro, which software is used to run this script?
@@ameeraarif Jupyter Notebook provided by IPython, pip library, numpy, pandas, xlsxwriter
Just completed this amazing journey
Thanks for all the efforts!
Literally just started building a trading bot, timing couldn't have been any better, thanks!
RUclipss algorithm is your friend.
You should thank the cookies for that
I’m also curious how your trading is going
I worked on it for a full month using the Interactive Brokers API in python and watched a bunch of videos focusing on day trading indicators etc to try and code the bot. In the end, I hadn't accounted for the commission the brokerages take (about $5 per trade which adds up especially if you're not trading with a lot) and this eventually started discouraging me a little. If anyone does run into the same problem, I would look into leverage which is riskier but easier to overcome tis problem. Risk analysis is also crucial (setting stop losses, take profits).
@@kenjohnson512 so how is it going bigboy?
Awesome course. Thank you, hugely helpful!
This is awesome! I followed every part of the course and finished all three projects successfully. Thank you so much for providing great quality content!
ah fellow Asian..
How do you feel like this course has affected your knowledge of algo-trading a month later? Have you been deploying any of the strategies used?
How where your results?
guys it is a simplified approach that allows to learn to program in python in a playful way. with this course you can try to make the analysis more relevant by integrating other indicators idk
where did he has shown how to install dependencies and cloning plz help
I am really enjoying learning these techniques and am impressed with your style. Thank you for encouraging others to learn!
I'm running into an error in the momentum calculations around percentileofscore implementation, around 2h 30m mark in your video. I am executing the following:
for row in hqm_dataframe.index:
for time_period in time_periods:
change_col = f'{time_period} Price Return'
percentile_col = f'{time_period} Return Percentile'
hqm_dataframe.loc[row, percentile_col] = score(hqm_dataframe[change_col], hqm_dataframe.loc[row, change_col])
and get the error:
TypeError: '
I found a solution to this question in one of the comments. Another user had the same issue and they resolved it by adding this line before the for loops I mentioned:
hqm_dataframe.dropna(inplace=True)
This drops any rows/columns that return N/A as part of f''{time_period} Price Return' column in the hqm_dataframe
Note that there are a few pops in Nick's sound at the beginning, but these go away completely once he starts building projects at the 17:20 mark.
great, was worried it was my system 😅
Hi, I am still a beginner, where do we go to install the dependencies... the video skips that part.. thanks
@@blockdisney3676 I think you can start with this anaconda
ruclips.net/video/YJC6ldI3hWk/видео.html
@@contantino_mm Thank you so much!
@@contantino_mm And where to start without this anaconda?
There were a few questions I had while watching (I am a noob so sorry if any of them sound dumb):
- How do you deploy that? I imagine you wouldn't want it running in Jupyter Notebook.
- Is the accuracy-loss of floats safe to ignore?
- While the concept of high-quality vs low-quality momentum makes sense, why is momentum a good metric? Isn't it likely that since you rely on data from a year ago, your model identify stocks that have already seen their growth and won't grow more? (Like making you buy at high, instead of low)
- Would it be sensible instead of calculating a score to sort all the columns based on different return_percentiles?
- Also, when do you know to sell? I know the premise was that we have a team of traders to whom we send our excel files, but while they'd be able to know how many stocks to buy based on our position and the filtering we did, how do they know when to sell?
Amazing video! Thank you so much for sharing this! Still going through it, but there's an observation I need to make for any beginners that might stumble upon this video:
To create a new column with values based on another column(s), using a for loop to iterate through rows and make calculations (as shown in 2:31:11) will be EXTREMELY inefficient on larger datasets. It is better to use Pandas' `apply()` method paired with a `lambda` function, as follows:
```
# create list with time periods
time_period = ['One-Year', 'Six-Month', 'Three-Month', 'One-Month']
# create new fields (or update existing ones) based on previous list
# and calculate percentile based on Price Return
for i in time_period:
change_col = f'{i} Price Return'
percentile_col = f'{i} Return Percentile'
hqm_dataframe[percentile_col] = hqm_dataframe.apply(
lambda x: score(hqm_dataframe[change_col], x[change_col]), axis=1)
```
Same method can be applied to calculate HQM Score using Pandas' `mean()` method:
```
hqm_dataframe['HQM Score'] = hqm_dataframe.apply(
lambda x:
pd.Series([x[f'{time_period[0]} Return Percentile'], x[f'{time_period[1]} Return Percentile'],
x[f'{time_period[2]} Return Percentile'], x[f'{time_period[3]} Return Percentile']
]).mean(), axis=1)
```
Hey Alex! just checking, are you still able to use it? Did they deprecate the sandbox api?
@@Jon-bk2bw Not sure if it's working at the moment, but it did a few weeks ago when I tested it! If anything, you can use Yahoo Finance API instead
@@Jon-bk2bw I had the same question and I think they did on 21st August 2021 but I'm not sure how come users were able to work with it until a couple of months ago.
Thank you so much for this tutorial! It has been a great distraction from my regular school work. Keep em coming! :)
About to take a college course on this topic next semester, so glad this exists!
which course and college? could you please refer to a link?
Waste of time.
@@irvin1241 Im sure most colleges that have a business school within them will have a course or two on python or R that’s centered around finance. The course Im taking next semester is a minicourse to fill out my credit requirement so it’s probably not as in depth as youre looking for. The best place to start would probably be just googling the “python finance college course” and seeing what colleges offer that you’d be willing to pay for. Of course, if youre good at learning things on your own I cant recommend Kaggle enough. Its free and has alot of stuff that you can pull from. Good luck!
@@gomes8335 I dont think its a waste of time, as Im not looking to be a full time programmer. I pay for 18 credits per semester, and using a few of them to focus on my interests outside of my major seems pretty worthwhile to me. I’m not a good self-learner like some people and my background in coding isnt that strong. I think it’ll be a cool way to combine a minor interest with my other education and see if I can really take it anywhere. As always, the internet is truely the best teacher when it comes to programming, but not everyone is suited to that path :)
@@VERY_TALL_MAN Good on you. I wish you luck :)
The fun thing about pandas, is that you can work with whole columns without loops, e.g:
final_dataframe['Number of Shares to Buy'] = final_dataframe.apply(
lambda row: math.floor(position_size / row.Price), axis=1)
Best tutorial. New in python and Stock is foreign to me. I actually can have stock conversation, now. Thanks
Loving the materials!
I might add that Pandas and Numpy actually has builtin methods better suited than math and scipy libraries for some operations.
For example, numpy.floor() can be used instead of math.floor()
final_dataframe['Number of Shares to Buy'] = np.floor(position_size/final_dataframe['Price'])
or df.rank() can be used instead of scipy.stats.percentileofscore()
for time_period in time_periods:
hqm_dataframe[f'{time_period} Return Percentile'] = hqm_dataframe[f'{time_period} Price Return'].rank(method='max', pct=True).mul(100)
Oh, man have a nice day, u made mine easier
thanks for .rank
@@adiletdaniyarov5737 HQM Scores from momentum strategy,
momentum_percentiles = [f'{time_period} Return Percentile' for time_period in time_periods]
hqm_dataframe['HQM Score'] = hqm_dataframe[momentum_percentiles].mean(axis=1)
RV Score from valuations:
rv_dataframe['RV Score'] = rv_dataframe[[i for i in metrics.values()]].mean(axis=1)
:)
That's awesome, thanks you
Bro thank you so much I was getting float errors.
I knew free code camp would have a legit course on this.
Thanks for the training. I learned alot!
awesome!Fix some problems:Before Calculating Momentum Percentiles,Execute the statement first to clean up invalid data.hqm_dataframe.fillna(value=0, inplace=True)
Great thing I have this on my notif
Awesome video, thanks very much! For the next time, please put the video of the educator not on top of the slides or if in a corner where it does not superimpose on the content.
This tutorial is freakin' awesome!!
Your presentation of this information is fantastic and easy to pay attention to, thank you you're awesome.
Shouldn't stocks with negative P/E and P/B ratio be removed when you are calculating the RV Score? I see that it was done when you take the simpler approach and only use the P/E ratio. Thanks in advance :)
Thanks for the video Nick!
If i wanted to launch a career in trading along with my coding skills, which videos do you suggest me to watch in order to have a solid background on the subject
thanks in advance
This video is insanely helpful. Thank you so much.
A great detailed starter course to learn some Python. As a licensed commodity trading advisor (and a software engineer), I would never advise someone to trade this though.
?
Could you explain your opinion or suggest anything?
52:39 If you stuck with appending series to a dataframe like me, this method(.append) returns a new object, so if you want to see the result, you should assign final_dataframe to new variable and then print it.
new_dataframe = final_dataframe.append(pd.Series(
[
symbol,
price,
market_cap,
'N/A'
],
index=my_columns
),
ignore_index=True
)
print(new_dataframe.to_markdown())
TY! assigning the new variable got it to print and installing tabulate
Perfect! Exactly what I got stuck on.
Just coming across this video now. Make sure, for the first project, you get an up to date list of all stocks found on the S&P 500. At this time there were 4 stocks that no longer are under the listed tickers with the provided csv file.
I'm very grateful for your video. I'm below a beginner, basically illiterate in computer knowledge, but you made it easy. Thank you, thank you!
where did he has shown how to install dependencies and cloning plz help
Great video! Was really helpful. Just 1 small comment/question:
In the 3rd value investing project, you might have missed out filtering out the stocks with negative pe ratio/pb ratio/ ev_ebitda ratio / ev_gp ratio before calculating the percentile. The algo is currently rewarding such stocks with negative ratios which is incorrect, right?
Rather than counting zeros, an easier way to input large numbers with many zeros is to use e-notation:
3e3 (which equals 3 * 10^3, or 3000)
10e6 (equals 10 * 10^6, or 10000000)
8e100 (equals 8 * 10^100, or 80000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000)
Also an easy way to input small numbers, e.g. 3e-3
Are you trying to input my bank accout balance?
@@Zaenil Yes, for a withdrawal, but all I keep getting from your bank are 🤣emojis. What's up with that?
Just as a safety precaution since you never want to trust someone to know what to do you can use a while loop for the input validation:
invalid_input = True
while invalid_input:
try:
portfolio_size = input("Enter the value of our portfolio:")
val = float(portfolio_size)
print(val)
invalid_input = False
except Exception as exception:
print("You must enter an integer")
BONUS: In this course a free addictive Python course for beginners / intermediate programmers is included, with clear clarifications about everything you may need ♥
As sandbox is discontinued I'm using the public key provided by IEX within the test period. I'm still getting 403 error. Any ideas?
Just completed the course. Thank you!
Your teaching style really touched me. It's so much connected.
What I like about your tutorial is that one doesn’t miss out if they start with the third section.
Also than you didn’t cut out the bugs and struggle helps people to understand that these stuff belong to a programmers life.
Another positive point is that you try to clarify specific financial terms.
Overall you did a great job here.
One point of improvement is your face cam position on the screen.
where did he has shown how to install dependencies and cloning plz help
Furthermore, you could try to
1) follow the approach of Ang (2009) and exclude firms with the lowest 5% of market valuation (and not only prices
One of the most challenging courses on flying helicopter, preflight. Checking some of these instruction will benefit you greatly. Former Airforce.
Hello, for people getting to the Key Error for the stocks, the solve is to delete the 4 stocks that have been delisted from the S&P 500:
code example:
stocks = stocks[~stocks['Ticker'].isin(['DISCA', 'HFC','VIAC','WLTW'])]
thanks!
Thanks!
i cant find this Needd Heeellppp!!!
Section 2: Course Configuration & API Basics
How to Install Python
Cloning The Repository & Installing Our Dependencies
Jupyter Notebook Basics
The Basics of API Requests
@deqiang tian Thanks and it worked! My steps are a little different 1) I downloaded "Anaconda" and 2) In the Anaconda Prompt I type "pip install panda"
"pip install numpy"
"pip install jupyterlab".
OMG. He talks so fast, but explains concepts so darn clearly..... 👍👍👍👍👍👍👍👍👍👍👍👍👍👍👍👍👍👍👍
playback 0.75 might help a bit
@@omgitzcoola I love fast speakers:) just amazed by him.
this was great, many thanks Nick
Really liked the tutorial! There are a few places where you're working on code that's blocked by the video of your face. I could generally extrapolate from what you were saying, but wanted to let you know
hi,
can anyone please help me figure out where actually is a dependency section in this course?
Thank you
I know nothing about stock trading. I might as well learn it so I can watch this course!!
Incredible course. Thank you very much!
Well my next five hours are gonna be well spent 👌🏿👌🏿
Chowa cross, now I think we are actually on the same page.
@@noelcollins2355 hahaha dude getting deep into algo trading man ..it's dope asf
just 5? I would estimate it would take me more than 10 hours!
2x
Sandbox testing has been deprecated. Is there any alternate free data api that we can use in this project?
Excellent explained and walking through learning session! Great thanks and hope there will be more sessions from you!
Unbelievable! Yall doing sucha great! Thank you very much! God bless yah family!
Great respect from Kazakhstan!
It seems IEX has closed access to the sandbox mode unless you have an account which starts at $49/mth. (30 day free trial).
Never wrote one line of code in my life and I feel ready to go work for google. Amazing!
You clicked on so many lightbulbs for me with this video. Thank you so much.
Hey man thanks for this video. Really great work here.
at 2:28:12 While implementing the percentiles, I noticed that the data I am receiving has None type in it. So the scipy.stats module throws an error as it cant compare the None type instances. For now I am replacing it with 0. Just for people who face this error you can just add this right before you assign your return percentiles
for row in hqm_dataframe.index:
for time_period in time_periods:
if hqm_dataframe.loc[row,f'{time_period} return'] == None:
hqm_dataframe.loc[row,f'{time_period} return']=0
I actually did a hqm_dataframe.dropna(inplace=True), as this values are from stocks that are probably not available anymore using the API
Hey thans guys, you all saved the day :)
@@fcoreas Thank You this fixed it for me
@@fcoreas This worked perfect for me! thank you.
@@fcoreas eyvallah francisco
'Citadel'- Cue in Fortunate Son and flashbacks of Vietnam with my WSB brothers.
You should do one that actually goes through candle data for several stock tickers. That's what seems to be lacking everywhere I go. People always trade on just 1 stock ticker when they go through candle data and do TA strategies. Would be nice to have a tutorial on going through candle data maybe even intraday, and buying specific stocks from a long list of tickers using a % of available funds, or % of portfolio.
couldn't you implement what was taught in this to accomplish that?
I’m doing the cs50 Python course atm, but will follow this roadmap when I’m done… thanks !
Yeah, its very unfortunate and discouraging to those who are learning the fact that section 2 is missing from the actual video even thou it is mentioned on all the syllabus for this course that I came across. Several people on the comments are complaining about this and I mean... You could at least edit the video description with basic instructions or with a link to where we can find the missing part of the video. After all that is critical information for those who wants to follow along with your course. Thanks anyway
you know, Renaissance technologies is something I pass by on my way to almost anywhere and I never knew what they were until now. That's pretty cool
they dnt use python
@@pajeetsingh What do they use?
@@ayansamal2440 A quick Google shows C++
@@goober-ll1wx If you can place order 100 times a minute, it's still not illegal. They place 1000 times a min. It's not illegal. It's calculus of day trading, more like minute or even seconds trading.
@@goober-ll1wx Well he worked for CIA. That should settle your doubt.
This was amazing! Thank you so much for this video!
Thank you very much for the videos. I wish you had covered the execution of a trade. That is an area that I often ran into a trouble. I hope we will have a series that cover execution and an actual end-to-end pipeline based on an actual exchage.
@Afshin Zarechian I now primarily use tradingview or tradestation for writing my scripts and backtesting. They have a program called pinescript which is easy to learn. They both have built in backtesting module.
Awesome tutorial, watched the whole thing through to the end. For the next video tutorial, do you want to show us how to write a trading bot that will automate day trading based on these algorithms?
Bro u intrested to learn more about algo trading?
And the fact that this is free re establishes my belief in humanity
There is no such thing as free
Not free for long
I know absolutely nothing about Python… But this is the first video that seems to clearly explain what my brain just can’t seem to wrap itself around. When I designed the investment strategy that I use, it was simple math, well, it was math… But to suddenly take the algo and transfer it into a language that I am not familiar with, seems like the worst possible prostitution of my time…
I get it, learn Python then write code and be happy ever after but, but, but, why can’t I do that in English? I can read medical journals on just about any topics, all I need is a dictionary and some extra time, I don’t need to mortgage a significant portion of my time on this planet just to learn something that will be old info by the time I’m done learning it… There’s gotta be a practical way to do that in English! Heck in French or Spanish or Russian or any languages! Geese this is so frustrating!!!!!!!
Amazing tutorial! Thanks so much
Very confused on how i missed this but where does he list the dependencies i need for this? I watched the intro and am new to python, I'm not too sure what environment he is using or how I install everything I need for this course
Hello.
If you're installing this on Windows 10, make sure to install the version from the python website and not from the microsoft store, since both of them install python in different locations. As for the dependencies, use the command prompt (don't use the python IDLE), and install the dependencies by typing:
pip install
Replace dependency-name for the following modules:
numpy
jupyter
pandas
requests
xlsxwriter
Hope this helps you continue
@@tetsuo79 Sir, I can see that you are a kind man. Can you please help me with getting into jupyter notebook please. as I am facing some problem with it.It seems I need some password to access it.
Thanks u soo much sir
U provide free education
❤️❤️ Love from india
This is incredibly valuable info