7 replies to this topic

[ned]

Hi everyone -

I am a grad student working on a paper and I need a little help.
I am new to GIS, so go easy on me

I am trying to determine how census blocks map into precincts in one county: San Diego.
[For those of you who don't know, census blocks are generally smaller than the precincts, but there is often overlap (one census block might be split by a precinct line).]

I have the shape files for both the precincts and the census blocks (they are free to download at http://www.sangis.org/Download_GIS_Data.htm).

I would like to determine what percentage of each census lies in each precinct. So, basically, I want to overlap the maps and then figure out the areas of intersection (let me know if that does not make sense and I will clarify).

A few questions:
1) Is this possible?
2) I am interested in learning GIS - can I do this by myself or will it be too difficult (I have some programming skills but have never really used GIS...I do have access to arcgis)?
3) If I hired someone to do this, what is a reasonable rate? Does anyone know someone who would be interested?

Thanks for your help!
Ned

Edited by ned, 05 February 2009 - 03:24 PM.

[p-dub]

Hi Ned,

My name is Patrick, I'm in my last semester of an undergrad GIS degree. I'm very interested in economic geography (Minor in Econ), and I did some work with census tracts last semester, so your question caught my eye.

What you want to do is possible and a rather basic GIS problem. It is an "overlay" operation. If you have access to ArcGIS and want to research how to do it, search the terms: Intersect, Union and Identity. It sounds to me like an 'Identity' procedure would get you there. (I believe 'Identity' is an ArcGIS specific term, but any GIS should have a similar function)

I'm sure you've had some other responses, but I'm a starving student, I'd give it crack for cheap.

So why do you need this information, what is the significance of the overlapping areas? What is the question you are trying to answer? Just curious

Best regards,

Patrick Wild

patrickwild@gmail.com

[pfyfield]

Actually "Union" would be the way to go. This would give you polygons with the attributes of both input polygons. Run an area calculation (those attribute values won't change in the union overlay, and will be incorrect in the output layer) and you should have all the information you need. You'll need to do some table manipulation in excel or access to get the percentages.

This is a very straightforward procedure. The tools are all in the toolbox.

[MapMedia]

Hi Ned,

I would use intersect, because you want to know the intersection of blocks and precincts, calculate the new overlap areas, then calculate percentages.
You can send an email to christen.erichsen @ gmail dot com if you need help.

[Jimbo]

I would like to add the following comment that if this is a project anything to do with your education then it's best that you do your own work because what if you hired a GIS technican to do your work & your professor found out. What grade would you expect if you are caught cheating? So let me suggest to you that you terminate a call for bids on elance, the lowest offer is above $400, especially when the majority of providers only offer second rate work, so can you as a student afford that?

The best thing to do is do the work yourself, get the assistance you need, & learn. Now it's been awhile since I have used ArcGIS but I know rate away from a Manifold point of view that after the intersection is done it involves some SQL to create that cross tabulation.

[ned]

Hi everyone -

Let me first say thanks for the help. I definitely appreciate all of the ideas and contacted Patrick about getting the work done. Christen - I will very likely contact you in the coming weeks if I decide to take on this project myself.

I did want to clarify the situation to Hasdrubal:

I would like to add the following comment that if this is a project anything to do with your education then it's best that you do your own work because what if you hired a GIS technican to do your work & your professor found out. What grade would you expect if you are caught cheating?

I am a graduate student doing a PhD in Economics and am writing an econometric paper on voting behavior (that I intend to publish in an academic journal). I have never studied GIS and am not trying to cheat on a GIS assignment. It is common to use someone else's technical help on an academic paper like this, assuming that you do not take credit for that part of the analysis (which I will not). Just thought it was important to make that clear.

I guess the question is simply whether it is worth investing in knowing these GIS tools for future projects (so I don't have to rely on someone else's help in the future). I think I will take a look at ArcGIS, spend a day and see how far I get.

Thanks again for all the help, and I am still happy to listen if anyone has good ideas/tutorials to point to, or if someone knows someone who might be interested in working on this project.

Ned

[frax]

Ned, it is not that your school has a GIS lab or geography department? I am sure you could find a helpful lab manager or grad student there to assist you...

[gregsd]

... I think I will take a look at ArcGIS, spend a day and see how far I get.

Ned,

Do you have access to MapInfo software? If so, then you can use a simple SQL statement to return the results:

Select precincts.Name, census_blocks.Name, proportionoverlap(precincts.obj, census_blocks.obj) from precincts, census_blocks where precincts.Obj intersects census_blocks.Obj into Selection

The proportionoverlap() function will return the percentage overlap between 2 objects (census blocks and precincts). You wouldn't even need to convert the shapefiles, as MapInfo can open these directly as well!

HTH

Greg.