Open Source GIS (Fall 19)
Access our SQL here
This week, we used data downloaded from Resilience Academy, a Ramania Huria project, to analyze urban resilience in Dar Es Salaam. As one of the most mapped cities in the world, we are taking advantage of the wealth of qualitative and quantitaive, locally-collected and analyzed data to start creating our own models of urban resilience. I worked with a partner, and our research question concerned analyzing an aspect of flood risk, percentage of subward area in the floodplain, to the various subwards in Dar es Salaam. From downloading the data that interested us to writing SQL to producing our final interactive maps, below are the step-by-step processes for this lab.
(refer to the SQL to see specific field names and functions)
Try: think of ways in which storm drain location information could be used to benefit a deeper analysis of flood risk in Dar Es Salaam!
For our project, we performed a Quick OSM search for all OSM tags like “wetland” and added that to our map, making sure to first transform it to the correct Coordinate Reference System (CRS). We then added in identification columns (like ward name) to make a spatial joins with the subward layer possible. This is documented in the first few steps of our SQL code, accessible on this page.
Then it was time to itersect the wetlands layer with the subwards layer in order to determine in which subward the floodplain fell. This part was the trickiest for us to implement using SQL code and it can seem quite confusing, so we recommend using online tutorials and other resources to better understand the SQL functions at work here.
Next we calculated the area of subwards and then the area of the floodplain in each subward using the st_area function in SQL. It’s a fairly straightforward function which calculates the area of given geometries, which can then be put in either the subward or floodplain tables. For the later purposes of vizual representation/mapping, I recommend working on the Subwards layer. However, note how it is possible to pull information from other tables to use in SQL queries within another table, which is what we did for the final combination.
The very simple next step involved determining the percent of wetland area in each subward; we were interested in vizualizing this as a primary factor of risk in flooding.
Next we added information from the Drain Location layer for vizual analysis: this included scripting in a count of drains in each subward. However, as mentioned earlier, you could really take this component much further to benefit analysis, but that simply was not in the scope for our lab.
After we created a map of these steps, we also worked with interactive mapping tools this week, like Leaflet. This is a really cool program which allows you to vizually present your analysis and then post it to easily-accesible web pages to make sharing your work that much simpler. First, we prepared our QGIS layout to rendered into html code using the QGIS2Web plugin. Using that html code and editing it to display our resilience map how we’d like, we created an interactive web map that anyone can check out. Remember there are a lot of edits you can make to the html code, in case certain elements didn’t transfer correctly. In addition to that, there are numerous other Leaflet features you can write into the script to create your desired map output. Check out this web page for more information.
Take a look at our final map here
Our map helps display the inherent, inequitable distribution for flood risk in Dar es Salaam, by our one measure. It is important to note that this is not meant as a cumulative assesment for all Risk. If one wanted to try and create a more accurate metric of flood risk, many other factors would have to be considered: average income, rent/property value, emergency access routes, storm drain infrastructure, and the list goes on. We simply wanted to illustrate one of these variables, percentage of subward in the floodplain, to begin to vizualize community risk. As we can see, certain portions of the city experience higher levels of risk for flooding naturally, and due to location residents there will be disproportionately impacted by the flood. Accurate assessment of risk is important in large decision-making processes in urban planning. For example, when governmental structures are allocating funds to help in flood recovery, they should be well informed of the different levels of need within the city.