Seeing Across City Lines

In the course of one month, the StreetCred community created, validated, and enriched more than 25,000 places. These data represent both new places contributed by the community and partner data improved by users on the ground. In bringing both kinds of data together, we are able to get an even better view on the POIs in LA and NYC.

Bringing the Community to Partner Data

 

The StreetCred community enriched more than half of the ~15,000 data points imported from partners

The StreetCred community enriched more than half of the ~15,000 data points imported from partners in the course of the month-long competition. Our partners rated their confidence in their initial data as high, low, or likely out of business. This created a new opportunity for the StreetCred community to prune off bad points from otherwise good datasets, as well as to find the gems buried in murky or possibly outdated data.

 
 
imported data type.png

In the case of partner data with the highest suspected quality, nearly 60% of POIs were flagged for removal, highlighting some of the inherent challenges in maintaining reliable data without real world validation. Low quality partner data was culled by 85% by our users, enabling us to successfully pull out nearly 750 viable places. Even among the places thought to be out of business by our partners, users found more than 6% of these to be good points.

 
 

NYC and LA

By taking all of the data generated by the community -- both the newly created and enriched partner data -- we can start to see patterns in how these two cities compare to each other.

More than 90% of the places in the data are either shopping, dining, or miscellaneous services

As world cities, LA and NYC unsurprisingly offer up a huge diversity of services and activities. More than 90% of the places in the data are either shopping, dining, or miscellaneous services. One way to start teasing out differences in these cities is to examine the variety of places and services each have to offer, and how this is represented in StreetCred POI data.

POIs in LA are heavily skewed toward shopping, which makes up nearly half of all places in the city (meanwhile civic and education-related places make up <1% each). NYC has a more even distribution of types of places, with shopping, dining, and miscellaneous services making up equal shares.

 
 
categories.png
 

The cities differ not just in the kinds of places they are home to, but also in where those places are located. The spatial arrangement of places can tell us a lot about how people use the cities they inhabit and the services they can access. As a preliminary cut at comparing these patterns, we see stark qualitative differences in the mixtures of POIs between NYC and LA.

In NYC, our community generated dense clusters of points in lower Manhattan and the Bronx. Within these clusters, there is a rich mixture of all categories of places.

When we zoom in on downtown LA, the types of places seem to be more clearly divided by area. Shopping points (blue) are heavily concentrated in the Fashion District to the South, with dining (orange), services (green), and travel (red) to the North.

These city flyovers give a flavor for how we can use StreetCred data to explore what cities have to offer and how people use them. We’re looking forward to digging deeper into these patterns, and expanding the view as the community collects more data across the country!

chris shughrue