-
polled last 14 months from bom (http://www.bom.gov.au/climate/dwo/IDCJDW5002.latest.shtml) and writing script to strip out - now in 'weather.csv
-
ended up using data from form at http://www.bom.gov.au/climate/data/, possible could automate through file name. in file airportweather.csv
-
ftp only has current data and forecasts
- http://binalytics.wordpress.com/2012/06/18/quick-web-app-with-r-rook-and-ggplot2-for-a-stock-symbol-candle-chart/ using rook and ggplot
- better dynamic server - shiny
- http://www.rstudio.com/shiny/
- https://github.com/rstudio/shiny-server/wiki/Ubuntu-step-by-step-install-instructions - host locally.
- https://github.com/ramnathv/dashifyr - dashboarding framework using js plugins knob rcharts and shiny server - looks much harder to use than shinydash - need js
- https://github.com/trestletech/ShinyDash - another dashboard built on shiny server and js
- nned to install gridster js library somehow?? http://gridster.net/
- awesome example, exactly what we need - http://glimmer.rstudio.com/winstontest/shiny-jsdemo/
- use html slides w inline r - like sweave but nicer to look at. can include rCharts by saving as html then loading.. look at tester.html
- slides w day, season etc
- use slidyfy w embedded r charts - e.g. 'tester'
- rcharts - highcharts
- http://shopify.github.io/dashing/
- d3 - need to learn actual javascript of jquery
- intially export mail manually matching srearch: from -- eventually automate to scrape of entire outlook backup - pst.
- ubuntu cli readpst -S to individual plaintext files
- awk/cut/sed to strip out data from each email -- play with file 'sample_email.txt'
- ramp up time for new wash
- daily split/ weekday/weekend
- total number washes
- washes in OH
- percent premium wash
- time of washes - within oh or without
-
weather - can scrape from BOM, GSOD?? -- sunlight - threshold to categorical? -- precipitation - binary? -- temperature - threshold??
-
weekend vs weekday
-
public holiday
-
season - 2 or 4?? - categorical.
-
demographics - population,
-
shopping centre numbers>?? can get?
- day mix per wash in multibarchart
- season vs rain and sun - justify using climate data.