2. Locating the Stops - ScandinavianSection-UCLA/hGIS_ETK GitHub Wiki

An address locator can be compared to a GPS; it can be used to give a place name a location on a map. This works because each place in the address locator is connected to geographic coordinates that specifies the location that in turn can be matched with the place name with no geographical data. From http://danmarksstednavne.navneforskning.ku.dk/ about 200 000 place records with corresponding longitude and latitude could be downloaded into a common excel table. These could in turn be imported into ArcMap as points on a map.

2.1 Creating the address locator

There are some rules to think about when creating the Excel file that is going to be imported and used in ArcMap:

  1. Field headings can have a maximum of ten characters.

  2. The excel fields containing longitude and latitude need to be set to “Numeric”. By doing so, the coordinates with several decimals will be classified as “Double” (numeric) rather than “String” (text) in ArcMap.

The following steps were used to make sure the field are numeric (Excel 2013):

  1. Highlight one of the columns to be set to numeric.

  2. Under “Data”, click on “Text to Columns”

  3. Use the default settings – “Delimited” Next “Tab” Next “General”. In the last window, there is also an option to click “Advanced”. Do so, and chose “,” as the decimal separator. Leave the thousands separator as the default “.”

Once the field are organized, the file can be imported into ArcMap:

  1. The Excel table is added to the table of contents in ArcMap.

  2. Right-click on the Excel table > “Display XY data”.

  3. Specify the fields for the X and Y coordinates, with X corresponding to the longitude field and the Y to the latitude field. If the values from the Excel file would not have been registered as numeric, ArcMap would not recognize the longitude and latitude as coordinates, hence not allow for the data to be displayed on the map.

  4. The points displayed on the map are only part on an “event layer”. In order to generate an actual shapefile to work with from the imported data, right-click on the event layer in the “Table of Content” and choose to “Export” the data into an appropriate location. For eight out of the 200 000+ records, it was obvious that there had been an error when coordinates were registered because they were not located in Denmark. These were simply deleted. From the newly imported point layer, the address locator could be generated:

  5. Right-click on the folder you want the address locator in > “New” > “Address Locator”

  6. “Address Locator Style” needs to be specified, and in our case “General Gazetteer” is the most appropriate for our use and is described as “A gazetteer is a geographical dictionary or directory for place-names” (http://desktop.arcgis.com/en/arcmap/10.3/guide-books/geocoding/commonly-used-address-locator-styles.htm#ESRI_SECTION1_81054CF66AF8437A9E90CA580CE8496F ). The “Reference Data” is our new point shapefile with all the place names.

The address locator works as a search engine with the result from a place name search generating the points that corresponds to the search result. How precise the match needs to be to generate a result was left to default with the “Spelling sensitivity”, “Minimum candidate score” and “Minimum match score” at 80%. This can be changed for individual geocoding attempts. The adjustable spelling sensitivity is good since there has been some change in how a couple of letters are written in Danish. Now they will generate a match even if they do not match with the search to 100%.

2.2 Geocoding the place names

With the address locator generated, the Excel files with extracted data can be geocoded, meaning the stops Kristensen made can be put as points on the map.

  1. Add the excel file of a fieldtrip into the Table of Contents.

  2. Right-clicking on the sheet, there is an option to “Geocode Addresses”.

  3. The address locator to base the search on needs to be specified to “allPlaceNamesLocator” generated in the steps above (“see 2.1 creating the address locator”).

  4. The “Address table” (see below) is going to be the excel sheet with the place names, and the search in the address locator will be based on the field “Place” in the excel table (see 1. Place Name Extraction). The output shapefile will be named FT_XXX_stops.

The output shapefile will be the first draft of the stops Evald Tang Kristensen made, but many of the different places he visited will have many “matches” in the address locator since there is multiple locations with the same name in many cases. To get the likely stops he made, some work needs to be done in the “Interactive Rematch” (see below).

  1. Once the first draft shapefile has been added to the map, the user is asked if she wants to “Rematch” the results. The answer to this is yes. To access the interactive rematch after closing the window the first time, add the “Geocoding” toolbar by clicking “Customize” > “Toolbars”. Highlight the point shapefile with geocoding results and click on in the geocoding toolbar.

  2. Sort by the field “Sequence” by right-clicking on the field and choose “Sort Ascending” to go over the stops in chronological order.

  3. Highlighting one row at the time, each representing one stop, generates a varying amount matches that show up as ice blue points on the map. Once a stop is highlighted in the candidate window it is selected and show up as yellow and a little bit bigger on the map. Clicking “Match” makes the stop a permanent point in the shapefile.

  1. Tang Kristensen usually made his stops in a logical order. This means that the stops tend to be close to each other – especially if he walked. Using this logic, we found the most likely matches for each of his fieldtrips.

  2. Note that the matches can be re-done, and it is possible to go up and down the list of stops several times before settling for a final selection. Once finalized, the interactive window can be closed and the stops shapefile is finished.