Learning about Detroit by talking to every person isn’t possible - it’s just too hard and expensive. Luckily, we can use mathematical equations to figure out how many people we need to talk to in any area to get an general idea of what people think. But we had a problem: Detroit’s council districts are too big, and Detroit neighborhoods are too different from each other for this method to work easily. To fix this, the project team came up with “Neighborhood Zones”. These zones make sure our data is reportable while also respecting Detroit’s neighborhood boundaries. In this post, we’ll explain how we make these zones and why they help us get better information.
The Neighborhood Vitality Index is a tool to track progress and advocate for resources to make changes in Detroit’s neighborhoods. The index measures things Detroiters actually care about—including health, well-being, feelings of safety, and more—by including data from multiple sources, especially from Detroit residents. The data comes from traditional sources, like the Census Bureau, but also from a survey of Detroit residents where questions about individual neighborhood experiences can be asked to obtain more data about lived experiences within specific communities.
The best way to understand a community’s voice would be to conduct a census and speak with every single resident in that neighborhood. This is what the Census Bureau attempts every 10 years with the federal census. It’s extremely time consuming and expensive to identify, contact, to talk to, and record the responses of all those people. So annually, the Census Bureau gathers data from the American Community Survey (ACS), which asks a smaller group of people (also called a “sample”) to respond to a survey. After the survey responses are collected, additional mathematical equations are applied to get to an estimate of what the whole population’s response would be. To learn more about how data collection works, you can read more about how the pandemic impacted data collection. These equations also give us a margin of error, which describes how reliable an estimate might be (a more in depth description of margins of error by NVI partner D3’s Comparing Data: How to Use Reliability Measures in Real Life). We use these ACS data points in NVI for things like housing cost burden and income levels.
We recognized early on in the development of the NVI that it would be impossible to talk to every single resident of Detroit to understand how they feel about their community, so we had to come up with a way to collect a sample of Detroit residents that was evenly distributed across the city. We could not use neighborhood boundaries because Detroit neighborhoods vary a lot in size, making it difficult to compare between them. Detroit neighborhoods also vary a lot in what we call population density, meaning that in some areas there are a lot more people living close together, while in other areas there are fewer people living far apart. Since so many neighborhoods are small or have lower population density, it would have been almost as time consuming and expensive to survey enough people to create estimates at the neighborhood level as a full census. An obvious solution was to utilize Council Districts, however residents and CDOs we spoke with knew it wouldn't provide them with the detail they needed. So we created Neighborhood Zones as a compromise to capture more details for the CDOs while also grouping a large enough population for data collection.
Part of the neighborhood zone development process was to understand how community development organizations (CDOs) and residents identified their neighborhoods’ boundaries and come up with a strategy that balanced data integrity with the authenticity of those boundaries. What exactly is a neighborhood zone and why didn’t we just use the neighborhoods that Detroit residents already identify with?
Neighborhood Zones in NVI were created based on neighborhoods, and also keeping with the boundaries of Council Districts. With the input of residents and community organizations, we divided Council Districts into 3-4 geographic areas. These are each identified by a number and a letter, so, for example, Neighborhood Zone 3a is a zone located in Council District 3.
When people take the NVI survey, their address provides a specific point in the city and with our handy geographic database, we’re able to identify what Neighborhood Zone they live in. We take all the data in that zone and aggregate it to calculate the Neighborhood Zone’s average response. Aggregating data means that we take all of the individual data from people living in the neighborhood zone and group it together to get a better understanding of that specific area. How we group it together depends on the data, sometimes we show a range of responses, other times we use an average or median response.
For some other data points, we use other data sources like the American Community Survey. This data is only reported at the aggregated geographies so we have to approximate which geography belongs to which neighborhood zone. For NVI, we chose to use Census tracts, which are basically small building blocks within the Census Bureau's geographic areas. We use census tracts because they have more reliable estimates compared to some of the smaller geographies, like block groups. If you’re curious about what a census tract is, you can explore more about census geographies on State of the Child.
As one of the benefits to CDOs promoting the survey to their residents, we offer a special datacut of NVI data that matches the CDO’s boundaries. Since the NVI data is collected at the point level, the CDO boundaries are able to be replicated exactly. However, for the census tract data, we use the tracts that align most closely to the CDO’s boundaries. Sometimes this means the data represents a slightly smaller or larger area than the CDO. This is a limitation of the data we’re working with and we will do our best to help you understand the geographic scope of a dataset with a map highlighting the data boundaries.
And that’s how and why we created Neighborhood Zones. By breaking Detroit into smaller, manageable areas, we can report more accurate data that respects the unique character of each neighborhood. These zones help us use NVI data to understand more about different areas of Detroit without losing the big picture.
