food access in Amsterdam

Roboat is a 5 year, multi-million dollar research collaboration between 4 MIT labs, TU Delft, and Wageningen University studying the potential of autonomous boats (ro[botic]boats). As part of my work at the Senseable City Lab, I proposed a system of floating, autonomous markets that could supplement the robust network of markets already present in Amsterdam, highlighting the potential to tap into the greater region’s food production. Initial analysis of markets in Amsterdam showed a complex network of transportation and supply.

Using open source GIS software and publicly available data from Amsterdam’s web portal, I conducted a closest facility analysis in order to determine the distance one would have to walk (or bike) to the nearest supermarket. This analysis was run for cells measuring 100m2, allowing for a higher granularity in the results. Overall accessibility to supermarkets in Amsterdam is reasonable considering definitions of food deserts, however, I sought to discover if there were inequities in the ways supermarkets were distributed in the city. For that reason, the density of each cell was also calculated using the listings of residential addresses. These two measures were combined to find cells that were both dense and had a long distance to the nearest store, with the assumption that this meant there was another factor than demand at play. The map at right and key above shows these results-- Q3 denotes that the value is in the third quartile, or the top 25% of results.

Once it was determined if an area of Amsterdam was located in a vulnerable area where density was high, but access to a supermarket poor, I sought to determine if the difference could be explained by demographic data. A statistically significant difference was found using the share of non-native Dutch, meaning that the data showed a relationship between the ethnicity of an area and its access to a supermarket.

To test this hypothesis, I developed a model that controlled for both density and Average House Value (WOZ), factors that were considered proxies for spending power in order to isolate the effect ethnicity has on supermarket placement. However, it was found that for every percentage difference in the share of non-native Dutch, the distance to the nearest supermarket actually decreased, albiet by .17 meters. What this means is that a cell that is 63% non-native Dutch has 3 meters fewer to walk than one that has 39%--a very small difference, but a statistically significant one.

Considering, however, the massive amount of growth the city of Amsterdam is projected to see, there may yet be a place for Roboat to aid daily life. The region has hundreds of thousands of households planned and in exploration since 2015. The radial diagram above shows the households as they are distributed between zipcodes in the city. The chart below visualizes the proportion of these units that are located on the major waterway the IJ. Could this be the natural habitat for an autonomous water-borne service provider like Roboat?

As such, I imagine Roboat markets could function as individual stalls, able to dynamically appear on the canal side for Amsterdammers to collect crates of fresh produce or the Roboat markets could combine together to form larger, more typical markets, on the water or inside buildings. Rather than using autonomy to remove people from the system, the use of autonomous boats could bring more people together to purchase healthier food and engage the larger region.