-
@beeonaposy talking about 8 years of data science mistakes. Sharing mistakes helps with transparency, to help, and for growth. Mistakes count as experience. #rstudioconfPermalink On twitter.com
Mention
❤️ 2 Favorites Mood 0
-
@beeonaposy Types of mistakes: technical versus communication. Early on, mistakes tend to be technical, but communication mistakes are important too. #rstudioconfPermalink On twitter.com
-
@beeonaposy Built a model - but some variables really couldn't be used. Usual suspects in analysis: understanding data source; assumptions; filters; missing values (& didn't catch the last one). #rstudioconfPermalink On twitter.com
-
@beeonaposy Predicting patients who could fill in for cancellations; ran into time zone issues; lack of shared resources about data; institutional knowledge depleted as people left. #rstudioconfPermalink On twitter.com
-
@beeonaposy Communicating with business stakeholders; rolled out multiple features, but couldn't show impact to single features. Encountered artisanal data in Excel. #rstudioconfPermalink On twitter.com
-
@beeonaposy Rhetorical triangle - frame analysis for anyone. Three elements: speaker, audience (are they technical?), and context (are you delivering good news/bad news?). Can help with working together. #rstudioconfPermalink On twitter.com
-
@beeonaposy Make sure you understand the business problem and that you're doing it right. Frame analysis so that they make sense. Basic facts (who, what, how man?). Know where the data comes from #rstudioconfPermalink On twitter.com
-
@beeonaposy Pseudocode that lives next to the actual code is important. Data dictionaries, and code review. Core team meetings really important for data analysis. #rstudioconfOn twitter.com