What I have: I have a large data-set containing links (bus routes( (>2000) and their respective nodes(bus-stops) (>6000). The data set I received from the local bus operator is very messy. It contains multiple variations of names and multiple (nearby) locations for the same stop. There is no clear trend of these variations, so I cannot run a program (say, to merge all points within 50 m radius etc. Example in orange and yellow circle in the picture).

What I want to do: I want to use QGIS clean the data manually.

Is there a method/plugin that lets you select multiple stops and merge their locations either to one of their locations, or even a manually entered location?

I now have ArcGIS (10.2) also, so a solution in either QGIS or ArcGIS is fine. Here is the screenshot from ArcGIS.

enter image description here

The points highlighted in yellow the circle are actual locations of the bus-stops. I want to merge them into one data-point(consider it as a node), while retaining their attributes. (attributes contain information about which routes pass through the stop and what order is the bus stop in a particular route etc ).

So if these data points are like this:

  1. "Intersection xyz1 (Besides Public School)", lat1, long1, route 23, direction up
  2. "Intersection xyz2 (Besides Govt. Hospital)", lat2, long2, route 23, direction down
  3. "Intersection xyz3 (Besides General Pharmacy)", lat3, long3, route 87, direction up
  4. "Intersection xyz4 (Besides Open Area)", lat4, long4, route 87, direction down

They should be converted to:

  1. "Intersection xyz", lat, long, route 23, direction up
  2. "Intersection xyz", lat, long, route 23, direction down
  3. "Intersection xyz", lat, long, route 87, direction up
  4. "Intersection xyz", lat, long, route 87, direction down
  • Can you put your data in Postgis? It would be possible to use St_Snap to do what you need. Do you have a layer with routable network lines? – Alexandre Neto Dec 27 '13 at 10:07
  • I had never used PostGIS, until recently. So not very comfortable with it. I have basic SQL skills, so I will try using ST_Snap. Also, getting a route network layer from this data is the where I am headed. Once I have the clean data, that would be the next step. – Sameep Dec 27 '13 at 10:34
  • 1
    Some questions: (1) Can you rely on the names for the merge? That is, can you try to extract "Intersection xyz" and drop everything that looks like "<digit> (...)"? (2) What is your eventual goal? I'm curious about your request that the result should include the same number of points as the input, with the names and spatial locations harmonized. The idea of four points with the same name and location, but separate route numbers, directions, etc. is a little weird. What is your intention? – Lee Hachadoorian Dec 27 '13 at 22:12
  • (1) Not entirely. For two reasons (a) Some of the bus-stops ( especially towards the outskirts of the city) are generic names of villages. There are multiple villages with the similar (even same) scattered across the city. (b) Some bus-stops which represent the same node, have entirely different names. (2) The purpose of this exercise is to obtain a node-link type data-set. Hence ultimately after I harmonize the names and locations, I intend to have only one point for a bunch. This will comprise the node data-set. – Sameep Dec 28 '13 at 22:23
  • Please decide which of ArcGIS desktop and QGIS you wish to ask about in this question. Otherwise you are effectively asking two questions which goes against the Tour. – PolyGeo Jan 25 '17 at 10:06