[Maarten Buis]

Felix Kaysers #645 you can assign commands to F keys, see chapter 10 of the Users Manual.

[Abdullah Algarni]

I hope to see the following in Stata 18:

1. Stepwise regression with multilevel regression commands, specifically melogit
2. Improvement in tab1/2 command. For example, I can't change format of numbers and decimals
3. Adopt community coefplot command as default command, with further improvement (e.g., have interactive graphic control window)

[Bruce Weaver]

Hi Abdullah Algarni. Your 1st request in #647 reminded me of this FAQ on problems with stepwise (and related) selection methods. I'm sharing it in case you had not seen it before.

• https://www.stata.com/support/faqs/s...sion-problems/

I do agree that making -coefplot- an official Stata command would be nice.

Cheers,
Bruce

[Jonathan Afilalo]

It would be incredibly useful to be able to run contemporary tree based models like XGBoost in Stata. I know this can be done in the Python extension, but it would be more efficient and user-friendly to to it natively in Stata.

[wbuchanan]

Jonathan Afilalo I'm not sure I'd agree that a Stata native implementation of XGBoost would be more efficient. It should also be possible to use the existing plug-in capabilities in Stata to link directly to the underlying C/C++ code used by XGBoost without having to invoke the Python interpreter, but reprogramming a highly optimized algorithm in a higher-level language would likely negatively impact efficiency.

[Jonathan Afilalo]

wbuchanan You are surely right from a programmer's point of view, but I meant more efficient from a user's point of view! I can already picture the beloved Stata GUI interface for XGBoost's inputs and hyperparameters, would be so nice!

[Helen Strongman]

I would like to be able to use parquet file formats in Stata

[wbuchanan]

Helen Strongman I gave a talk at the 2020 Stata North America User's Group conference that addressed that: https://wbuchanan.github.io/stataCon...2020-readit/#/.
Jonathan Afilalo if someone writes a wrapper that passes the information to the underlying Python/C++ implementation it shouldn't be terribly difficult to generate the dialog box for it; that said, I've tried creating dialog boxes in the past without a ton of success.

[Abdullah Algarni]

Hi Abdullah Algarni. Your 1st request in #647 reminded me of this FAQ on problems with stepwise (and related) selection methods. I'm sharing it in case you had not seen it before.

• https://www.stata.com/support/faqs/s...sion-problems/

I do agree that making -coefplot- an official Stata command would be nice.

Cheers,
Bruce

Thank you Bruce Weaver, I totally agree.

[Rabih El Habta]

My wish-list for Stata 18 are:

1-Increase characters length for variable name more than 100+ characters.
2-Insert table into SQL using the variables' labels.
3-Working with APIs -without calling Python.
4-Frames frlink to have all merge not only matched.
5-To have dynamic visuals not only static images, and save images in other formats such as svg.
6-GCE modelling.

[Ali Atia]

My wish-list for Stata 18 are:

1-Increase characters length for variable name more than 100+ characters.
2-Insert table into SQL using the variables' labels.
3-Working with APIs -without calling Python.
4-Frames frlink to have all merge not only matched.
5-To have dynamic visuals not only static images, and save images in other formats such as svg.
6-GCE modelling.

3 -- I agree that increased support for APIs would be great. However, you can call command-line functions such as cURL to query APIs without any need to call Python.

5 -- As of Stata 14, you can export visuals to SVG.

[Orhun Gun]

Is there any updates when the Stata 18 is arriving?

[Michael Harris differentiated]

Pystata support for moving data to and from the new, faster, better Apache Arrow / pyarrow backend for DataFrames in pandas 2.0.

Pystata currently does not recognize the pandas pd.NA values as numeric, so I have to manually replace with np.nan which defeats some of the purpose. Pulling data back from Stata, there’s currently an extra step to recast from NumPy-backed to PyArrow-backed.

Also: I second the request for parquet file format.

[David Flood]

Ggplot "repel"-style labeling for overlapping labels in graphs, particularly for scatterplots

https://cran.r-project.org/web/packa...s/ggrepel.html
https://ggrepel.slowkow.com/articles/examples.html

(I know about mlabvpos but it never looks as nice as I want it to without a lot of manual tinkering. I usually end up using R.)

[Chen Samulsion]

Ggplot "repel"-style labeling for overlapping labels in graphs, particularly for scatterplots
(I know about mlabvpos but it never looks as nice as I want it to without a lot of manual tinkering. I usually end up using R.)

David Flood I second your proposal. Not only scatterplot need this technique, but also pieplot, scoreplot, loadingplot......