jsonpott

Data science account for @jasonpott. Interested in health informatics, emergency care, routinely collected data. | nurse | #rstats | opinions mine | rt != ❤️

[Jason Pott @jasonpott]

Seen lot's of good things about @duckdb. I'm struggling to find information about best practices in R. Unclear whether I should close it down at the end of a session? Can databases be maintained between sessions? The tutorials I've seen jump into analysis & miss out housekeeping.

[jsonpott @jsonpott]

@ChrisBeeley Code that cares

Care to code

Coding is caring

[jsonpott @jsonpott]

@brodriguesco Did you mean to use +

[jsonpott @jsonpott]

@neilgcurrie Filter means keep.

[Shreya Shankar @sh_reya]

Our understanding of MLOps is limited to a fragmented landscape of thought pieces, startup landing pages, & press releases. So we did interview study of ML engineers to understand common practices & challenges across organizations & applications: https://t.co/RLkNqz8XtG

[jsonpott @jsonpott]

@ssotelo001 @__AlexMonahan__ @neilgcurrie @pachadotdev I can then work with the duckdb using dbplyr.

[jsonpott @jsonpott]

@ssotelo001 @__AlexMonahan__ @neilgcurrie @pachadotdev I’ve just been building out a local duckdb to work with multiple “larger than RAM” tables. I need to join the datasets and filter to create the final dataset.

I was unable to read in the file to convert to parquet. Duckdb can import the delim file without reading into RAM.

[jsonpott @jsonpott]

@neilhimself that was rather good. I only read Sandman earlier this year (in anticipation of this release - was on my list for a while). It’s a rare thing to produce an adaptation that does the book justice. Congratulations.

[jsonpott @jsonpott]

@PhDemetri In healthcare we have literature databases which you can build up@a search query to describe your:
Population
Intervention
Comparison
Outcome

Then read titles and abstracts to decide whether to include in your review. Only read full text of relevant or related papers.

[NHS-R Community @NHSrCommunity]

NHS-R one-day workshop on Shiny, 15th September 2022.
Teaching you Shiny all the way from writing your very first application to writing a full fledged dashboard!
Led by @ChrisBeeley
Register here:
https://t.co/WUKLvpKN9v

[jsonpott @jsonpott]

@data_guy_ @rstatstweet #Rstats Tidy Tuesdays are a good way to work on the basics, if you can’t manage to using R/python in the work setting.

Go for it! The hours quickly add up and you will be astounded by the amount of progress you make.

[William D. Lopez @lopez_wd]

No one:
Me: 2022 Oscar Outfits as Public Health Graphs

[jsonpott @jsonpott]

@theimmunodoc I loath endnote, it’s very expensive and they are very anti consumer in their development. I’m really liking ReadCube papers which has some lovely features. They have academic discount as well, plus a free trial.

[jsonpott @jsonpott]

@_johnmackintosh @NHSrCommunity University enterprise but that’s for academic work not NHS.

[jsonpott @jsonpott]

@Emil_Hvitfeldt 🏃🏻‍♂️💨 Quick.Rproj

[jsonpott @jsonpott]

@JF_Grillo_Bio @rstatstweet Glad it’s of some use. I had help from a colleague who tried to understand what the plots presented and kept telling me they were too complicated until I got it right.

[jsonpott @jsonpott]

@JF_Grillo_Bio @rstatstweet I hear you. I found that reducing the complexity to evaluate specific pathways resulted in Sankey’s that were better at describing processes in my data. I also merged groups where possible (e.g. rare events to ‘ other’). Consider breaking out individual processes for narrative

[jsonpott @jsonpott]

@brodriguesco Nice , I’ve been doing this lately but haven’t looked at the knit expand function just using iwalk with custom function involving cat, print and ‘asis’