Giles Robinson

Data Scientist at Sheffield City Council, UK. Social Care data & demand, modelling, SQL Server, R & dataviz. When I'm not doing that I pick a big bass fiddle

[Giles Robinson @GilesRobinson1]

#30DayChartChallenge day 3 - historic. Got data about battles & when I told #ggplot to scatter long v lat it gave me this nice world map of course. Today I finally got extrafont to work & I'm learning plenty about annotation, colour & legend control. #dataviz

[Giles Robinson @GilesRobinson1]

@knrdblzk @prosser55 @rstats_tweets +1. Better not to overlay

[Giles Robinson @GilesRobinson1]

Mild selection bias at work here in that 100% of the data points in the "I do not wish to provoke" category are situations where the shark attacked anyway.

[Giles Robinson @GilesRobinson1]

#30DayChartChallenge day3 is flora & fauna. Found some #sharkattack data & enjoyed the provoked/unprovoked column so much I made this irresponsible #ggplot monstrosity. Seems it's the unprovoked attacks that'll get you killed. Learned about adding white space & element_markdown

[Giles Robinson @GilesRobinson1]

#30DayChartChallenge day 2 calls for a waffle and I had this Pokemon data already so here's a pokewaffle. Hex codes off the interweb to get the proper pokecolours. Couldn't get geom_waffle to work & faceting would be nice but hey it's Sunday I'm going to bed. #dataviz #Pokemon

[Giles Robinson @GilesRobinson1]

try & do something colourful tomorrow

[Giles Robinson @GilesRobinson1]

#30DayChartChallenge day 1 and I was driving today so ended up thinking about roads, getting some data about roads and coming up with this one quite crap statistic from it. #ggplot2 #dataviz #slightlycrapstatistics

[Giles Robinson @GilesRobinson1]

@markgannon @BBCLookNorth

[Giles Robinson @GilesRobinson1]

@BryonyShannon ...so it makes sense to look at those groups separately, and to do so you have to label the groups. Hadn't occurred to me that people may dislike being referred to by acronyms though, so I'll think more carefully when presenting my analysis! (2)

[Giles Robinson @GilesRobinson1]

@BryonyShannon Well, one way we understand systems is to break large groups into smaller categories. E.g. people with learning disabilities and older people are (mostly) 2 distinct groups with different needs, financial situations, long term prospects & demographic trends. (1)

[Giles Robinson @GilesRobinson1]

@DrChrisGibbons I think this captures the true spirit of Christmas

[Giles Robinson @GilesRobinson1]

@BryonyShannon Also client / patient / citizen / service user. What's the preferred term these days? I don't really like any of those. "People" is always good of course but I often need to distinguish between different groups of people: staff; ppl who access social care & ppl who don't

[Giles Robinson @GilesRobinson1]

@BryonyShannon Interested in responses to this as an analyst in Social Care. When presenting data you want to be as concise as possible so I say & use LD all the time! Rare that my audience is the public but still want to get the lingo right.

[Giles Robinson @GilesRobinson1]

@DrChrisGibbons

[Stig Abell @StigAbell]

Tier 1
Tier 2
Tier 3

[Giles Robinson @GilesRobinson1]

@WelshLassinShef Apologies for the public shaming! Only mention it because I just got a big felt desk mat and it's made my workstation feel really plush. Hoping it'll reduce glare off my very white desk also.

[Sheffield City Council @SheffCouncil]

This is the most up-to-date information we have on COVID in Sheffield. All data is accurate at time of publication but is subject to revision by Public Health England.

This week we have graphs!

[Giles Robinson @GilesRobinson1]

@DrChrisGibbons @Nigella_Lawson I love a rendang. It ticks all my curry boxes.

[Giles Robinson @GilesRobinson1]

@Berbesque deepliyar

[Giles Robinson @GilesRobinson1]

@DrChrisGibbons techincal term for this device is an embiggener