Stanford University began working with the City of San Jose, California, in the spring of 2017, as an opportunity for students to gain experience on real-world projects in the urban systems analytics space. After San Jose was chosen as one of the three pilot cities for the USA Sustainable Cities Initiative, the team focused on developing a digital dashboard to track the city’s progress towards the UN’s Sustainable Development Goals (SDGs). This virtual and interactive tool would collect, organize, and help visualize various sets of data to allow the public and local policy makers to better understand and address the SDGs.

One of the foremost challenges in this task was prioritizing exactly what features would be best to focus on for use by those in municipal government and the public. Considerations mainly included balancing the accuracy of our analysis with its complexity – so that the results are both insightful and understandable by those with varying degrees of data literacy. We decided the dashboard needed to serve two primary functions. The first task was to use a handful of visual tools that would most neatly organize the data to make it approachable and intuitive. And second: we wanted to format the data (or put it into certain units) in a way that would make the most sense to those who would actually use it. This demanded it be made tangible and accessible in order to actually motivate useful, directed action. Because of this high level purpose for the dashboard, much of the most novel data analytics work was directed towards SDG 13 (Climate Action). This offered the most potential for meaningful insight in a sector that the San Jose government had not previously measured in a robust fashion, and also provided an opportunity for individual citizens to use the information to inspire decisive changes in their daily lives.

Because of the necessity of effective visuals to communicate the data, and especially due to the desire to encourage real action with the dashboard, the geospatial representation of sustainability metrics could serve to “gamify” the information. The hope is that just by showing an individual citizen how his or her neighborhood, district, or city compares to other neighborhoods, districts, or cities (using shaded regions to indicate better or worse performance in any given category), he or she would be motivated just for the sake of being the best and/or not being the worst in any given category. We generally assumed that the more granular the spatial comparison for any given metric, the more ownership of the data would be imbued upon the individual, and likely the more impetus they would have to alter their habits. Beyond the mapping tool that would serve as the primary visual, the two other data organization strategies would use a graph and a data table, offering a simplified portrayal of how a metric might change over time or between regions. For instance, a line graph tracking progress towards a given SDG before or after certain policies, or a bar chart showing clear comparison between median or average values of regions within the city. These quick visuals might help to further motivate citizens or at least better reach those who only use the dashboard for brief amounts of time.

Core to the theoretical dashboard’s utility is the data’s accurate reflection of the goals and metrics they’re supposedly measuring. Because many of the SDGs are considerably abstract and do not lend themselves to a singular, intuitive, or comprehensive metric, essentially all goals require some form of proxy data. This is used to mimic the status of the system for each goal as soundly as possible. In many cases, we did this by using granular data from the US Census Bureau (American Community Survey with information collected annually) as the proxy to distribute more high level data down to the block group level. The block group represents the most granular scale of data available for most American cities. As an example, with the knowledge of how much electricity was used across the entire city of San Jose (provided by PG&E, the dominant electric utility), we used the annual amount of money spent on electricity bills at the most granular level to assign the proportional amount of electricity used in each block group.

The result of the work of the Stanford students can be found here, though it should be noted that this was by no means a final product. The project ultimately engaged several interested and interesting stakeholders and prompted the next step in development of the dashboard. Progress for this continued effort is expanded on here. Looking forward, far more work will be put into expanding the goals measured, and better researching data relationships to verify the accuracy of the data and proxies used to document progress towards each SDG.

Ensuring our data analysis can be interpreted by those outside of our team is equally critical. In that realm, more involved engagement at the local level, with the help of one of San Jose’s more active neighborhood districts, will let us responsibly refine the elements of the dashboard visual itself. If successful in these endeavors, the San Jose SDG Dashboard could stand as a far more trusted and frequently used tool for cities even beyond San Jose at multiple levels of the decision-making process.