You can verify that Stata is doing the math correctly and see a map of that route here and here, after you enter the lats and lons you get from traveltime3. The Haversine distance is much shorter at 1,544 miles, so is the Vincenty one at 1,546 miles. Geodist lat1 lon1 lat2 lon2, miles gen(distance_v)Īs you can see when you run the code, the travel distance (along Interstate 10) is ~1,838 miles, which you can verify with Google Maps. Geodist lat1 lon1 lat2 lon2, miles gen(distance_h) sphere Traveltime3, start(from) end(to) units(imperial) Gen from = "4905+Lakeway+Drive+College+Station,+Texas+77845" Here's an example that calculates the distance from the Bigfoot Discovery Museum in Felton, CA to Stata Corp HQ in College Station: clear As far as I know, there's no way to apply either method to an actual physical route to adjust for curvature. Both methods can be used in Stata using geodist from ssc (which has a great help file with references). The other method is to use the Vincenty equations (which is an slower, iterative process, but still assumes away terrain). This gives an ‘as-the-lazy-crow-flies’ distance between the points (ignoring the distance needed to go over any mountains or canyons on the way and assuming the earth is a sphere). The Haversine formula calculates the approximate great-circle distance between two points, which is the shortest distance over the earth’s surface. You can also calculate the length of the shortest curve between two points along the surface of a mathematical model of the earth. So given some maps, historical and real time data from various sources such cell phones, their algorithm picks the best route using a black box method. In another blog post, they link to the Wiki entry for Dijkstra's Algorithm when explaining how they pick the best route.
These factors down to a "cost" associated with each candidate routeĪnd display the route with the lowest "cost."īiking directions seem to adjust for elevation in picking the route (though that does not mean the distance is adjusted for it). The meaning of best is pretty fuzzy:ĭeciding which route is "best" involves balancing competing factors:ĭistance, travel time, number of turns, and many more.
Note that if you are trying to calculate the distances between all points, then you should take a look at geonear (also from SSC).Traveltime3 (and all its ancestors that work with previous versions of the Google API) uses Google's guess for the best driving/walking/biking distance along some viable route at a point in time. Geodist origin_lat origin_long destination_lat destination_long, gen(d)
#Stata ssc geodist how to
Using the data you provided in the comments, here's an example of how to use geodist (from SSC) to calculate the distances. Traveltime, start_x(origin_lat_1) start_y(origin_long_1) end_x(destination_lat_1) end_y(destination_long_1)įrom your comments in a previous thread, you have lat/lon coordinates and you are trying to calculate distances between those points. Mqtime, start_x(origin_long_1) start_y(origin_lat_1) end_x(destination_long_1) end_y(destination_lat_1)Įrror: type mismatch after 30 var but it also did not calc for the first 30 geodist can also calculate great-circle distances using the haversine formula. By default, geodist implements Vincentys (1975) formula to calculate distances on a reference ellipsoid. I already extracted the lat and long of my observations but now it does not work to calc the distances those a the commands: geodist calculates geographical distances by measuring the length of the shortest path between two points along the surface of a mathematical model of the earth. I try to calc distances with the mqtime and traveltime command but it does not work. Geocoding with Stata mqtime and travletime command
0 kommentar(er)
