• Microsoft acquired Revolution Analytics in January 2016

• Opportunities for development of open source technologies for sustainable planning

What is the PCT?

• An open source, publicly accessible web-based tool for assessing the geographical distribution and potential benefits of cycling potential nationwide

• Best illustrated in a demo: http://pct.bike/

• Where should we build for existing cyclists?
• New cyclists in the medium term?
• Long-term strategy?
• Along which routes to create the strategic joined-up network?
• Where should cycling interventions be prioritised around public transport infrastructure?

• Build and test a prototype model
• Identify 'desire lines' of greatest potential
• Make the tool scalable nationally
• Create a website that will make the Propensity to Cycle Tool a publicly accessible resource

Version 1 - nationwide (V1 launch: June 2016)

• Route-allocated hilliness, network layer (complete)
• Include health outcomes (HEAT)
• National-level results (Anna Goodman)
• V1.5 - smaller (LSOA) zones (Jan 2017)
• Training

Version 2 - local deployment

• Include non-commute trips
• Compatibility with Local Authority trip demand models
• Micro-level analysis (V 2.2)

• Case studies of use (e.g. Manchester, Kent, Yorkshire)
• Method for identifying severance
• Case study along HS2 route
• 'Hackathons' to stimulate the tool's development

pkgs = c(
  "stplanr", # transport data handling
  "sp", # package for spatial data
  "tmap", # package for mapping
  "deldir", # voronoi polygons
  "readr", # fast read/write
  "rgeos", # gis functions
  "dplyr" # data analysis
)
lapply(pkgs, library, character.only = T)

## [[1]]
##  [1] "stplanr"   "sp"        "knitr"     "stats"     "graphics" 
##  [6] "grDevices" "utils"     "datasets"  "methods"   "base"     
## 
## [[2]]
##  [1] "stplanr"   "sp"        "knitr"     "stats"     "graphics" 
##  [6] "grDevices" "utils"     "datasets"  "methods"   "base"     
## 
## [[3]]
##  [1] "tmap"      "stplanr"   "sp"        "knitr"     "stats"    
##  [6] "graphics"  "grDevices" "utils"     "datasets"  "methods"  
## [11] "base"     
## 
## [[4]]
##  [1] "deldir"    "tmap"      "stplanr"   "sp"        "knitr"    
##  [6] "stats"     "graphics"  "grDevices" "utils"     "datasets" 
## [11] "methods"   "base"     
## 
## [[5]]
##  [1] "readr"     "deldir"    "tmap"      "stplanr"   "sp"       
##  [6] "knitr"     "stats"     "graphics"  "grDevices" "utils"    
## [11] "datasets"  "methods"   "base"     
## 
## [[6]]
##  [1] "rgeos"     "readr"     "deldir"    "tmap"      "stplanr"  
##  [6] "sp"        "knitr"     "stats"     "graphics"  "grDevices"
## [11] "utils"     "datasets"  "methods"   "base"     
## 
## [[7]]
##  [1] "dplyr"     "rgeos"     "readr"     "deldir"    "tmap"     
##  [6] "stplanr"   "sp"        "knitr"     "stats"     "graphics" 
## [11] "grDevices" "utils"     "datasets"  "methods"   "base"

u = "http://www.dft.gov.uk/NaPTAN/snapshot/NaPTANcsv.zip"
zf = file.path(tempdir(), "NaPTANcsv.zip")
download.file(url = u, destfile = zf)
unzip(zf, exdir = tempdir())
stations = read_csv(file = file.path(tempdir(), "stops.csv"))
rail = filter(stations, StopType == "RSE")
saveRDS(rail, "data/rail.Rds")

# A tibble: 6 x 43
      ATCOCode NaptanCode PlateCode CleardownCode                CommonName CommonNameLang ShortCommonName ShortCommonNameLang Landmark
         <chr>      <chr>     <chr>         <chr>                     <chr>          <chr>           <chr>               <chr>    <chr>
1 0100BRP90087    bstgtaw                                  Brunswick Street             en                                             
2 0100BRP90088    bstgtjp                                   Brigstocke Road             en                                             
3 0100BRP90089    bstgtpj                                   Brigstocke Road             en                                             
4 0100BRP90090    bstgwgm                                    Denbigh Street             en                                             
5 0100BRP90091    bstgwta                                    Denbigh Street             en                                             
6 0100CLFDOWN0                                    Clifton Down Rail Station             en                                             
# ... with 34 more variables: LandmarkLang <chr>, Street <chr>, StreetLang <chr>, Crossing <chr>, CrossingLang <chr>, Indicator <chr>,
#   IndicatorLang <chr>, Bearing <chr>, NptgLocalityCode <chr>, LocalityName <chr>, ParentLocalityName <chr>,
#   GrandParentLocalityName <chr>, Town <chr>, TownLang <chr>, Suburb <chr>, SuburbLang <chr>, LocalityCentre <int>, GridType <chr>,
#   Easting <int>, Northing <int>, Longitude <dbl>, Latitude <dbl>, StopType <chr>, BusStopType <chr>, TimingStatus <chr>,
#   DefaultWaitTime <chr>, Notes <chr>, NotesLang <chr>, AdministrativeAreaCode <chr>, CreationDateTime <time>,
#   ModificationDateTime <time>, RevisionNumber <int>, Modification <chr>, Status <chr>

rail = readRDS("data/rail.Rds")
coordinates(rail) = ~Longitude+Latitude
plot(rail)

For more info: see this infographic on usage stats.

z = readRDS("../pct-data/cambridgeshire/z.Rds")
qtm(z) +
  qtm(rail, bubble.size = 0.3)

• Various ways of doing this
• Voronoi polygons is most visual:

rail_cam = readRDS("data/rail_cam.Rds")
voronai = deldir(rail_cam$Easting, rail_cam$Northing)
plot(voronai) 

for(i in 1:nrow(cents)){
  o = cents[i,]
  d = rail_cam[mat[i],]
  od = sbind(SpatialPoints(o), SpatialPoints(d))
  if(i == 1)
      L = SpatialLines(list(Lines(list(Line(coordinates(od))),"X"))) else
        L = sbind(L, SpatialLines(list(Lines(list(Line(coordinates(od))),"X"))))
}
ld = SpatialLinesDataFrame(L, cents@data["rail"], match.ID = F)

Using @robinlovelace tool (stplanr, modified) to review @cyclestreets journeys from their API, with @qgis. pic.twitter.com/t17LebZFJu

— Matt Turner (@MattTurnerSheff) January 31, 2016

• Building cycle paths to stations (layer in the PCT)
• Identify where there's a short-fall of cycle parking?
• Identify target areas to incentivise to cycle?
• Provision of cycles at 'arrival station'

• Find if rail is a viable option for non-rail trips
  • Trains may alread by at capacity (e.g. Ely)
  • Many destinations are not close to rail stations
• Find if nearest station is suitable per rail commuter - People may not go to the nearest station - TransportAPI or Google's Distance Matrix API can help
• Explore travel options upon arrival at 'destination station'
• Get data on cycle facilities at stations

• Email: r.lovelace at leeds.ac.uk
• Twitter: @robinlovelace
• Thanks for listening