Andy Pryke

I'm delighted to announce I started work at Prodigy this week 😃

I just love this! “I have students who are low-performing in my class as well as gifted students. In Prodigy, my low-performing students can compete with my high-performing students without tension or negative consequences, which causes them to perform better.”

@joranelias Sounds awful! Out of interest, what is your platform?

@FiveThirtyEight did you notice that fox news seem to have disabled their "probability meters" showing forecasts from their statistical model. Or maybe they just broke... foxnews.com/elections/2020…

@d_olivaw @olga_mie @useR2020muc Thanks!

Is there a recording of the #useR2020 Tackling Climate Change breakout panel? I'm afraid I missed it. @olga_mie @useR2020muc

Great that #useR2020 is now online. Missing the beers & chat & meeting/catching up with so many interesting people! user2020.r-project.org/program/agenda/ (note: more talks under the "program" drop down)

@StatStas Sounds closer to "Tragedy of the Commons" to me en.wikipedia.org/wiki/Tragedy_o…

Lockdown tip: Don't forget to update your LinkedIn with new in-demand skills. 😊

@ben_millwood @jamestplunkett @N_Damschke @tomaspueyo @Austen @paulg The link @ridsab re-posted works for me: docs.google.com/spreadsheets/u… and here's a link to a copy I made: docs.google.com/spreadsheets/d… If these don't work, let me know.

@jamestplunkett @AndyPryke @N_Damschke @tomaspueyo @Austen @paulg The original file is seeing a lot of traffic and it seems like Google are limiting how many people can view and copy it. If you have a working copy of it, can you consider making yours available for others as well?

When should your employees start working from home? Nobody made data-driven decisions, so I built a model. It uses data to calculate true spread & the likelihood your employees are infected #coronavirus @Austen @paulg docs.google.com/spreadsheets/d…

@ridsab @tomaspueyo @Austen @paulg That also works for me

@tomaspueyo @Austen @paulg Use this link he mentioned in his article to copy the model: docs.google.com/spreadsheets/u…

@N_Damschke @tomaspueyo @Austen @paulg Thanks. Following your suggestion, I opened the model in the Google Sheets phone app, then closed it, then from the list of recently opened file I could make a copy I can edit.

@AndyPryke @tomaspueyo @Austen @paulg For me it worked via the App for Google Tables/Spreadsheats. Then it doesn't open in the browser but the app and you can make a copy.

Great article on coronavirus risks with model to predict for your region. Charts clearly explain expected growth & why the official numbers lag behind the actual numbers. @tomaspueyo/coronavirus-act-today-or-people-will-die-f4d3d9cd99ca" target="_blank" rel="nofollow noopener">medium.com/@tomaspueyo/co…

Very interesting explanation of how to determine disease spread from virus genomes

Good visualisation of how the Wuhan Coronavirus compares with other diseases. Key takeaways: less fatal than some "big name" diseases, but might spread faster Source: nytimes.com/interactive/20…

@CaliCoyote e.g. we could switch in a python/scikit learn ML component, and have also experimented with using H2O autoML for model building.

@CaliCoyote We use: Relational DBs for feature engineering, audit & result storage; R for modelling, data/model auditing & report generation (rmarkdown); Jenkins for orchestration; .net with web interface for configuration UI. Interfaces are well specified - easy to switch components.

We have a highly automated data science system which builds, checks and deploys 100s of models (binary prediction, recommender etc) for clients each week. What open source or commercial options there are for this data science use case, or parts of it? linkedin.com/posts/andypryk…

@ETFOeducators can you give some advice for parents regarding picket lines? Our kids have full day daycare available in school. Does the picket line intend parents not to cross to use daycare?