Thank you Dorothy, Shawn, Chris, Justin and Donna for your quick responses. I now have a direction to begin. My questions and your answers are summarized below. Hello GISers, I have a few basic ArcView questions. If a manual exists to explain the process asked below, please let me know. I need this information before I can approach our college's computer service staff. First, I want to import our college's student demographic data into ArcView 3.2 (may be upgrading to 8.1). According to our CS, this data can be saved in delimited text in dbase, not Access or other formats. Second, I want to join this student data with the upcoming Census 2000 data, so that we can map "who" and "where" our students come from. Third, I understand that the database needs to have a primary key field, our students must remain anonymous-will social security numbers be legal and useable, or must we assign each record a new number? I do not yet know how to import this data using a key value. Once the database is imported into ArcView, what is the best field to use to join the student data with census data? The two entities will match on State, County, may be City, probably not addresses? Of course to assign each student to a census block or tract would be ideal. I need to know how to have tract, block, or city, or county specific 2000 census data joined to each student record. As I mentioned, if there is a textbook or manual available, please let me know. Thank you for your help, from a GISer 101. Hi, you can connect an Access database to AV using an OBDC connection and then doing a SQL connect within AV. I read (but can not verify) that AV files with SQL connects are more crash prone. Before I learned about the SQL connect I was simply exporting out the database from Access in a txt file and importing it into AV. As for assigning student addresses to Census blocks this can be done, you would need a layer with Census blocks (may be downloaded from the ESRI site) and another layer with the street addresses. Using the street address layer you would geocode the student addresses - I believe you would then assign them to the blocks using a spatial aggregation. The social security question is best addressed by your University policies. A pretty good AV manual is "Inside ArcView GIS" 3rd Edition by Scott Hutchinson and Larry Daniel printed by OnWord Press. I believe I got my copy from Amazon.com. Hope this at least provides you with some ideas. Dorothy Luckie Hi there Nancy, I work with census information everyday and have a few pointers for you. First, you can work with access files, and in some cases a hyperlink can be directed at the access files from AV3.2. Also, If the students have a student ID number that would be the best thing to do. If you used that ID and their address or zip code you could geocode them into the census geography. If you have more specific questions feel free to get in touch and I help as best I can. Shawn Shawn Blaesing-Thompson Washington State Senate GIS Analyst for Redistricting (360) 786-7641 :) Nancy, you definately need to join a tract or block group number to each student record to accomplish your correlation. You'll probably need to start by geocoding the students using their home addresses. You'll then need to get the census geography if you haven't yet (I recommend geographynetwork.com). Then, use the geoprocessing wizard to do an "assign spatial data" to the point shape file of the geocoded students. As you know, both the student point file and the census geography must reside in the same map projection before doing this. The spatial join will asssign a unique fips code from each tract or block to each student record. We've done similar things for public school districts here in Washington state. Hope this works for you too. Chris Hansen Delimited text or dBase is fine in ArcView 3.2. You just add a table and set the file type to delimited text. The data doesn't need a primary key field it is nice to have I would recommend stripping off the social # after you import the data into ArcView that way if you have a problem with any records you have a good identifier for finding the problem. But if you have a first name / last name that is probably enough and I would definitely take off the social. When you bring in delimited text any extra quotes make a difference. With the census data I would recommend Geocoding to the address but this is probably out of the question because then you would need street files for the entire country/ service area. The best thing I think would be to do a join based on zip code the census data in 2000 comes with a zip code boundary file. You may then want to do a summarize to find the number of students per zip code etc. As far as a manual Good LUCK Justin O. Niebuhr GIS Analyst St Charles County Government 201 North Second St Suite 310 St Charles, MO 63301 Telephone: 636-949-7900 Ext. 3441 Charlie Thomsen, an instructor at American River College in Sacramento, CA, has done some work like this with their students. I am sure he would be happy to give you some pointers. Also, ESRI has lots of links to colleges that use their software, you could email some of the contacts for ideas from them. Here is Charlie's info: thomsec@exi.arc.losrios.cc.ca.us http://ic.arc.losrios.cc.ca.us/~thomsec/ A few of my thoughts (background with federal and state government (including IRS)): You will probably want to geocode the student addresses, then do a spatial join if you want to look at census demographics related to each student. This takes the data from the polygon (census data) and assigns it to each point (student) within the polygon. Or you can sum by census shape and display and compare your data that way. You can use ArcView to geocode if you have a street base file to match against. Or you can have a third party geocode the addresses for you for a fee. Etak is one company I have heard of that does this. You could use the SSN as long as you don't release the SSN with any data. If you are the only user who will be dealing with the database it wouldn't be a problem. But if you have students making maps from the data, or would have the data on a public server, you should strip off the SSNs. You can use an ID number of some sort that can be assigned by your computer staff. However, realize that the data is still confidential because of the addresses (and names of course if they are used). You will need to find out what the college's disclosure policy is, and go from there. If the data will be used by students, consider summing the data by census tract or block group after it is geocoded. That way the identifying information -- SSNs, addresses, etc. -- will be removed. The dbase format is fine, that is what we generally use to bring data in to AV. If you don't want to geocode right away, you could bring the dbf in to AV, sum on the ZIP code fields, then join to a ZIP code shapefile. This won't be quite as useful, even though Census will be providing some information by ZIP, because ZIP codes change and aren't really accurate shapes (they are postal routes with a polygon drawn around them for GIS purposes -- actually they can intersect and do other strange things -- your data/analysis is not as good organized by ZIP Code -- but it could be a good first step). Hope this helps. Don't hesitate to contact Charlie, I am sure he would be happy to help you. Donna Bahls Research Program Specialist Labor Market Information Division California Employment Development Department (916) 262-2566 fax: (916) 262-2350 lmid.dbahls@edd.ca.gov Nancy Miller Acquisitions/Circulation Hinckley Library, Northwest College Powell, WY 82435 307-754-6167 FAX 307-754-6700 millern@nwc.cc.wy.us