Kampala Urban Poor

Analysis of water consumption in Kampala - slums vs affluent areas

Consumption data was obtained from National Water and Sewerage Corporation (NWSC) representing consumption per meter for a period of 2001 - 2003 and then 2009 - 2010 and then 2014. Data represents water meters at the various locations in the greater Kampala and Wakiso area as well as the consumption of water per meter per month during these specified years.

Amount of water consumed is measured in units. Each unit represents 1 cubic meter (1 m3) of water or 1000 Liters of water.

Data Selection

Data provided by NWSC for consumption was not complete. Figure 1 shows a graph of consumption vs period. For this analysis we selected the consumption in three months of 2014 (July, Sept, Oct), where the data seemed most complete. Data selected was averaged to obtain the average number of units consumed at each meter on an average month.

data selection
NWSC also categorizes meters in 4 general categories: domestic, public standpipe, institution or government and commercial. We selected only domestic and public standpipe meters for this analysis.

Analysis

We are interested in quantifying the difference in consumption between the urban poor (slums) and affluent areas in Kampala city. We quantify consumption per area based on the population of the area and the number of meters available in that area. Particularly we are interested in answering these questions:
  1. How does the number of meters in slum areas vary with the number in other areas?
  2. How does consumption vary across the different areas ?
  3. What is the population to meter ratio in slums vs affluent areas ?
  4. What is the consumption to population ratio in slums vs affluent areas

Population

We obtained population data from WorldPop - an organization whose purpose is to provide reliable population distribution data. For Uganda the data we obtained were UN estimates of population in Kampala averaged over a grid of 100 x 100 meters. We used 2015 estimates of the Uganda 2002 census estimated to match the UN population density estimates.
grid
The figure shows estimates of consumption averaged over a grid of 100 x 100 to match the population data. The slum areas are highlighted with red boundaries. However there is no marked difference between the slums and the other areas. Part of the reason for this is the estimate of the population is likely to be wrong for the slums because the population grows at a higher rate in slums. We thus sort to estimate the population of the slum areas.

Estimation of population in slums

To estimate the population of slum areas we digitized all the slum areas manually and estimated the size of a household in the slums. grid A total of 11 slums in Kampala were digitized including Mulago, Kitintale, Katanga, Kawempe, Namuwongo-soweto, Kosovo, Kifumbira, Kabowa, Luzira, Nakulabye and Naguru-godown. For comparison we also selected 5 affluent parishes in Kampala including Kololo I, Nakasero III, Nsambya Housing Estate, Makindye I and Kyambogo. Each house in the slum is estimated to harbor 5 people. We use the estimated area of a single house in the slum and the total area of the slum and the average number of people in each house to come up with the slum population.

Comparison across different regions in Kampala.

Analysis of consumption was hierarchical in nature. We compared the 5 different divisions of Kampala analyzing the consumption per division weighted by population as well as the number of meters located in each division. Analysis was further broken down to analysis of the different parishes in Kampala. More specific metrics were measured including number of units of water consumed per person per month and the number of people served by a water meter in each area. The last piece of analysis was a comparison of the slum areas vs selected affluent areas in Kampala. Similar parameters were compared.