Total Pageviews

Sunday, 6 November 2016

Tutorial: How to Map Ethnic Distributions?


Have you wondered how to do multivariable map of ethnic distribution? It has baffled me for a year or more as I was guessing on how atlases and magazines manage to produce highly detailed linguistic or ethnic maps of countries.
Ethnic Map of Chicago (Penn State University)
Recently, I just found the way to do so. In Malaysia, obtaining micro level data (at census collection district level or lowest level population divisible unit like U.S or Australia) is impossible. The closest lowest level demographic data I could obtain was electoral polling district level. It is not a complete representation on the ground but a very good sample set.

In this blog post, I will show you we can imitate those multivariable ethnic maps.

  • Excel Part
    1. Obtain Excel data on population data of the region you want to explore 

      Watch this video below on key things you need to consider in ensuring your Excel data is ready for a join in ArcMap. 
    2. As you can see the above Excel example, each row represents a polling district level data (for this instance). However, in order to produce those ethnicity maps as shown above, we should only consider the Majority ethnicity for the polling district level
    3. For this case, we should only consider percentage figures. 
    4. The question is how can I choose the column header (ethnicity) of the maximum row value of percentage (percentage of selected ethnicity). Refer to this article to implement -https://www.extendoffice.com/documents/excel/2547-excel-find-highest-value-in-a-row-and-return-column-title.html 
    5. Once we determined the majority ethnic group per polling district level, we need to classify them of varying percentage brackets (i.e. 30%-50%, 50%-70%, Above 70%). Two additional columns are to be added.
    6. For the first column, we will figure the percentage of majority ethnic group. We use the MAX function and highlight the attribute values (Refer here : http://www.accountingweb.com/technology/excel/identifying-largest-and-smallest-values-in-an-excel-list)
    7. Once completed, the second column with categorization is added. Using filters, group your data and label the data correctly.
    8. Prior to exporting to ArcMap or QGIS, copy and paste values only all the figures. This will remove all formulas as GIS system will not accept them
    9. Save the cleansed polling district data in a .csv format. Watch the following video for recap or clarification of the steps above. 

  • ArcGIS part

    1. In ArcMap, bring two datasets on the map using Add Data: .csv file of Population Figures and spatial data
    2. It is absolutely important that the file format and value of Identifiers of two datasets are identical. I couldn't stress about this more because the attribute joins depends on commonality of the values and format
    3. Right click the spatial dataset and choose Joins. Nominate the .csv file and choose the unique identifier. Run the Join
    4. Ensure all the features are joined. Identify which ones fail and rectify them either in spatial or csv datasets
    5. Once the Join is successful, go to Properties
    6. In Properties, proceed to Symbology and choose categories. The field for Symbology visualization would be predominant ethnic group percentage classification.
    7. Symbolize them with appropriate colour
    8. Click Ok to apply the colour scheme
    9. Check whether the desired map matches to your liking. Watch this video for recap or clarification of the steps above. 
Happy emulating my example!