The Analyst is a set of optional features in OTP 1.x used for transport analysis. These features were removed from OTP 2.0. This vignette will explain how to enable and use these features.
The Analyst adds the following features to OTP:
The analyst is an optional feature in OTP 1.x and can be loaded in
# Get OTP and the Demo Data in the normal way library(opentripplanner) path_data <- file.path(tempdir(), "OTP") dir.create(path_data) path_otp <- otp_dl_jar(version = "1.5.0") #Must use OTP 1.x otp_dl_demo(path_data) log1 <- otp_build_graph(otp = path_otp, dir = path_data) # Setup OTP with the analyst and pointsets enabled log2 <- otp_setup(otp = path_otp, dir = path_data, analyst = TRUE, pointsets = TRUE) otpcon <- otp_connect(timezone = "Europe/London")
You can see the analyst is working in the web UI.
You will have noticed that when we started OTP we also enabled the Pointsets feature. Pointsets are
.csv files of latitude/longitude coordinates that OTP can route to. PointSets can be created from SF objects using the
We will use the LSOA points from the Advanced Features Vignette to create a PointSet. We will also add a column of data called jobs to analyse. For any numerical values in a PointSet OTP can provide the count and sum of that value based on travel criteria. In this case we might want to know how many jobs can be accessed within a 30 minute drive of a given location.
otp_pointset() function takes the
lsoa object and creates a PointSet that can be accessed by OTP using the name “lsoa”.
A surface evaluates the travel times from a given point to all locations within 120 minutes in one minute increments. So it quite similar to an isochrone.
The fist step is to make a surface from a given location using
otp_make_surface function returns a list containing the surface ID and the parameters used to create the surface. You can use this object in other functions that support surfaces.
Once the surface has been made we can evaluate the travel times to our pointset.
otp_surface function returns a list of two objects. The first object
data is a data frame summarising the sum and count of any values in the pointset in one minute increments from the
fromPlace specified by the
surfaceid object. The second object
times is a vector of travel times in seconds to each of the locations in the pointset. NA values are returned if OTP was unable to find a route to the point or if the point is more than 120 minutes from the
If you are not using the sum and count features you can specify
get_data = FALSE to get just the travel times.
We can visualise the travel times using the
Notice that there are a few locations where OTP has failed to find a travel time. It is likely that point is too far from the road network to route, or the
mode does not allow access (e.g. driving up a pedestrianised road). You could try modifying your point locations slightly to allow OTP to find a valid route.
We can use surfaces as a more efficient way to produce a travel time matrix than using
otp_plan. A special function
otp_traveltime creates batches of surfaces and pointsets and summarises the results as a travel time matrix.
The function will return the travel times in seconds between the fromPlaces and toPlaces as a data frame. With columns names from
fromID and row names from
toID. Any failed routes will return and NA, as will any greater than 120 minutes.
The surfaces can also be used to produce isochrones from a surface.
otp_surface_isochrone returns a raster image with the value of the travel time in minutes. You can view the raster using the