Public Group active 1 day, 17 hours ago

GIS / Mapping Working Group

The Geographic Information Systems (GIS) / Mapping working group is a network of CUNY students, faculty and staff who are interested in sharing methods and techniques, and finding support from others about ways GIS can be used to further research and teaching.

The GIS/Mapping working group is part of a GC Digital Initiatives program designed to create collaborative communities of Digital Fellows, CUNY-wide graduate students, staff, and faculty to meet regularly and share their areas of interest. The working groups provide a sustained, supportive environment to learn new skills, share familiar skills, and collaborate with both the Digital Fellows and the CUNY digital community.

If you are using Geographic Information Systems or other mapping technologies in your teaching and/or research, or if you are interested in mapping your data, or using GIS technology to analyze/visualize your data, we invite you to join the GIS/Mapping working group.

Peruse our mapping resource bank here: https://commons.gc.cuny.edu/groups/gis-working-group/docs/gis-mapping-resources/

For the Spring 2024 semester, the GIS/Mapping working group will meet in the Digital Scholarship Lab, Room 7414, every other Tuesday from 2-4 p.m. Check out our event calendar for the specific meeting dates. Please stop by!

Spatial Join advice

  • Hi Everyone, Im not sure the forum is for individual requests of help, and if that is the case just let me know and I apologize in advance. In any case, I am asking for help and advice and I don’t know where else to ask or read by now.

    I wanted to ask if Olivia or Javier or anyone in the group could help me with a very frustrating issue on data management for my dissertation dataset:

    I am trying to join points (crime events) to a shapefile with city-blocks (polygons) of some Latin American in different countries (by IRB instructions some of the cities in my sample can’t be named, but you’ll notice) if you see it, especially Javier ;-)).

    So: in one of the cities I lose many of the points. They don’t join to the blocks (polygons). The polygons are separated as they have streets (polylines), and ‘splinters’ (errors) due to bad digitalization process. The polygons representing the city-blocks are very important because they contain a bunch of aggregated socio-economic data that will be precisely the quid of my dissertation. I should also add that many of the points fall out of the polygons in these streets, splinters, but I want to join them to the nearest polygon. These are the ones not being joined.

    So far nothing works, not with intersect join, or closest, or closest geodesic or any other type of match. I have to do this constantly as Im still getting data, and I will also join points of infrastructure like banks, churches, malls and schools.

    But be through a workshop or in person: can somebody please help me with this? This is not even central to my analysis, just the data management and Im totally stuck after days trying with ArcMap and Qgis.

    My emails are [email protected] and [email protected] and my phone is 201-706-1589.

    Thanks in advance for any help,

    David Topel Mujica

Viewing 4 replies - 1 through 4 (of 4 total)
  • Hello David,

    In this case it sounds like you have some bad polygon geometry. If you have polygons that are not properly enclosed they’re saved as line segments that lack area. An intersection between points and this layer will yield no result, as the points have no area they can fall within.

    In QGIS you can try running Vector – Geometry Tools – Check Validity on the layer to see how many problems there are. In the Processing toolbox under Vector geometry there is a fix geometry tool that can fix small errors but it can be hit and miss. ArcGIS has similar tools. I’ve had the best success fixing geometry using PostGIS.

    Other possible problems would include: trying to do an intersection between points and polys where you’ve plotted points but haven’t converted them to a layer (shapefile or geopackage) prior to running intersection, or a mismatch in coordinate systems between the point and poly layer. But if you are only having problems with your one city layer and the others work fine, it’s likely a geometry problem.

    This semester my GIS Lab at Baruch is open Wed-Thu-Fri. Hours and contact info are posted below. I am *not* in the office this week (Feb 18-22), but my lab assistant is in this Thu-Fri.

    https://www.baruch.cuny.edu/confluence/display/geoportal/GIS+Lab

    Best – Frank

    Hi David!

    There’s something that might be able to work out for you, once you’ve checked the validity of the polygons. Have you tried the QGIS plugin called NNJoin? It lets you join two layers based on nearest neighbor.

    With this plugin, you could join points to a polygon layer, although the caveat is that I think it will only take data from one point, not sure if it aggregates numeric data. The alternative is to do the opposite: join polygons using NNJoin to the points, so each point will have the data of the nearest city-polygon. Then you can aggregate the data yourself and using the unique key attribute of the city-polygons (name, or id number) you can then do a regular spatial join of the points to city-polygons, based on that key attribute.

    Besides the GIS Lab that Frank mentioned above, you can also come to GC Digital Fellows Office Hours, the schedule is here.

    I’ll be there next Monday 27 February, 3-5pm, would be happy to talk about your project there in person.

    Best,

    Javier

    Hi David,

    Just checking in to see how things went… were you able to come up with a solution?

    Best,

    Lisa

    Thank you Lisa for following up, and Javier and Frank.

    I have done it the way Javier explained, joined polygons to points and then used the unique ID to rejoin to the polygons. The features Frank identifies are true also, so thank you, but I still can’t repair them yet. The add-on NNjoin is something I didnt try yet. For the workaround Javier described, I didn’t need apparently to repair the geometry. Im still interested in doing that, so I will email Frank aside to go to his lab at the time of his convenience and show them all the layers I joined and the ones I am still working on to do future joins.
    My idea is to finish a dataset complete enough that I can export to Stata where I am more comfortable.

    Thank you again for your support and advice!

    david

Viewing 4 replies - 1 through 4 (of 4 total)

You must be logged in to reply to this topic.