Data Gathering

    Now  we continue building on our Frac Sand Mining project by collecting the necessary data needed to run a complex model.

     For our project, there was a few steps we had to take before we could geocode our data. First we had to gather our data from a wide variety of sources and put them into a single database table. We gathered data for Trempealeau country off of the county website. Luckily for us Trempealeau county keep track of all the sand mines located within its boundaries. You can see the data and website at the link listed here.  http://www.tremplocounty.com/landrecords/  Next we found more data at
http://www.wisconsinwatch.org/2012/07/22/map-frac-sand-july-2012/
Here is where the main body of our data came from. They had collected mine locations throughout all of west central Wisconsin. These two websites is where our wide range of data came from.
The problem with our data is that there was no sort of organization to it. We had to normalize our data in some way, shape, or form. We formed groups and each took a chunk of data. The major problem was the address where in multiple forms, so by using aerial photographs and trying to identify locations of mines off of aerial photographs, we went through our data to see if it was accurate and cleaned up some of the records. This cleaning process elimated some mines that were listed twice, but also removed some of the mines that were not identifable because they had no address.

     We then combined all of our data and connected to a GIS server to run our geocoder. We had to have common fields so we broke our data down into addresses, city, state, and zip code to help the geocoding process out. After entering in the necessary data, the geocoder used entered data to place a point on the map.We ran into some problems because we did not have addresses for all mines so the geocoder place the points in the center of the city that was listed. This is where using our aerial photographs helped because we are able to search to see if we can identify our mines. Another problem were some of the mines were listed twice so we had to remove some of the mines that were redundant.

     Are results were as followed: we were able to id 74 mines, and we were unable to id 25 mines. This is not including the Trempealeau county data which added some more mines. We still were able to showcase a wide range of mines in our data and had high number of mines found so we can proceed to our next step. There still could be some slight error because not all mines are opertaional at the time of the aerial photographs, so we have to rely heavily on the fact that the addresses on the websites were accruate to begin with.

Here is a map of our mines