Data science applications with TabPy/R
Вставка
- Опубліковано 16 тра 2024
- In this session, we will learn how to deploy connections between Tableau and its advanced analytics external services. We will then utilize these connections to develop powerful and flexible data science applications that make the power of advanced statistics and machine learning available to business users and dashboard consumers at all levels.
Hi Nathan, Thank you so much for your awesome presenting.
Thanks so much for this!... It provided me with many useful insights!... Great work!....
Where are the datasets used for this exercise? Just curious how one could connect with them.
Thanks Nathan, would you mind sharing the PCA workbook?
Hi, great video but I have one question. At 29:02 I saw Selected Variable in green, what is in that calculation?
Hi Nathan, thanks for the video. I am trying dynamic forecast in R. Everything works, except that the forecast doesn't happen for future dates. Can you share the workbook please?
Hi Nathan, thanks for the great presentation. How would I access the data sets and notebooks used in this presentation?
Connect Tableau to your data source + write the code in R or Python + launch Rserve (for R) or TabPy (for Python) and add the arguments in Tableau Desktop. The whole process is well documented on Tableau ...
hey Nathan, thanks for your sharing this is good. i know how to use tabpy locally but we have a case where we need to enable the tabpy analytics extension in tableau online (after we publish the dashboard to tableau online, we still can use the tabpy to run the python script we built in the tableau dashboard) and we want to host that tabpy server in a google cloud virtual machine, but we don't know how to develop the connection so that we can use tabby in tableau online. what is the URL to use in order to build the analytics extension? can you help us or is there any tutorial documents i can refer to? thanks
Really great webinar and superbly motivates the significance of TabPy! I am a data science engineer and would like to request if I can get a link or request the data set from the implementation? I am trying to teach the Python/Tableau implementation to students and you clearly and concisely explain it. I just want to show them a similar approach using your data set.
My email is pijcabasag@gmail.com.
Awesome video Nathan!
Can you share with me the dataset. Thank you.
@@trinhhainam He does share it. However, you need to write him an email.
@@PatrickBateman12420 Can you give me his email. Thanks.
@@trinhhainam Ask him in the post. I prefer he gives it to you.
Hi Nathan, Thanks very much for presenting and sharing the work. Its awesome and will going to be helpful in my project as well. I would appreciate if you can share me the workbooks.
He does, but you need to send him an email ...
Hi Nathan, Thanks very much for presenting and sharing the work. Great Seminar. I would appreciate if you can share me the workbook
Anyone know where he obtained the car.csv dataset?
Thanks in advance!
This is fantastic and has really helped me get started, but I'm a bit confused. At 29:02, I see PCA Component 2 on the rows shelf. What is in that calculation?
@@nathanmannheimer871 thank you! I'll give this another go and get it working 😊
@Ryan, just a little hint from painful personal experience: PCA only works if the problem has a linear shape. If the problem is non-linear, PCA can lead you astray. In this case, with non-linearity, better use e.g. Random Forest feature importance.
@@PatrickBateman12420 thank you for the added input. I managed to get this working (thank you again to Nathan). And my company finally adopted SAS Miner so I can better utilize more appropriate analyses like random forest, etc.
@@@ryanjayp . Awesome. Good luck!
Hi, by any chance did Nathan Mannheimer share the tbwx (or workbook) file with you? And if so could you share it with me? Thanks for any help
where i can find the data?
Is there a video that explains how we automate R to Tableau connection? I mean we surely cannot start RStudio always, run Rserve and then open Tableau and test connection. Am looking for such resources.
Correct, no need to fire up RStudio for production. You need only to connect Rserve with Tableau locally or on your server. For the Multiple Linear Regression, there's an alternative. Get the Intercept and Coefficients from R (or Python) and add them in Tableau's Calculated Field. By doing so, you don't need to connect to Rserve (or TabPy for Python). Of course, if there are updates, you need to update them manually in Tableau. In general, though, regression updates are rare. This is a neat trick and I'm surprised it's not well known. Let me know if you want to see an example.
@@PatrickBateman12420 , thanks Franco for the response. If you are planning to do a video on how to automate the R-Serve connection, that would help. Please me know if its possible... however, I use optimization algorithms to do some calculations and hence was interested to know the process to automate the R-Tableau connection.
@@MrKaddu Hmmm, I don't see the problem. Once you've launched Rserve (locally or on a server), you're "automated" in the sense that new R code is immediately visible in your Tableau workbook.
Correction, I wrote gibberish. You need to run RStudio or R (locally or on a server) in order to connect R + Tableau.
Hi, is there any chance I can get the tbwx file of this presentation? I can´t see any email so I don´t know how to ask for it, Thanks!
appreciate if you can tell me where can i download this workbook, thanks so much
Hi, by any chance did Nathan Mannheimer share the tbwx (or workbook) file with you? And if so could you share it with me? Thanks for any help
Can you share with me the dataset. Thank you.
Great talk but depressing how there is a one-to-one row relationship in data passed. I am sure there are some design reasons on Tableau's end but this really limits functionality for certain applications without having to do nonsense like moving a point forward in time.