Clustr helps Cities and Citizens Engage with SeeClickFix Data

Today is National Day of Civic Hacking! To join in, SeeClickFix is hosting a wide range of engineers, designers, civic leaders and entrepreneurs at our headquarters in New Haven. This group will come together to address the needs of our community, using SeeClickFix tools and APIs to build applications that solve city problems. A local developer, Allan Visochek, has already got us started on the right foot, with a really interesting data visualization and exploration tool, build on SeeClickFix.

The app — appropriately called Clustr — brings together groups of related issues so that cities can visualize problem areas more easily. These issues can then be browsed in clusters or visualized within a heat map. If you want to drill down into a specific issue, the app serves up a link back to as well.

Here is a graffiti cluster that I found right outside SeeClickFix HQ:

Cluster Image

I then drilled into the cluster where I was able to see specific graffiti instances and click through to the associated pages on

Cluster Image

This is a super intuitive and revealing way to browse issues throughout the City. Right away, I found myself moving from cluster to cluster, drilling into groups of issues — graffiti, potholes, illegal dumping, etc. It's really helpful to be able to click through to SeeClickFix as well.

From there, Cluster allows you to rank the each street based the length of a particular street, the number of active issues on that street, the total number of active issues for the given category, and the collective length of all streets in the given city where issues have been reported. You can see a detailed of the algorithm at

Cluster Road Rank

Just one example of the value of Cluster is what we found while exploring graffiti in Oakland. There appears to be a continuous ring of graffiti around Lake Merrit. It is really rather striking.

Oakland Clustr

To build the app, Allan brought together a team including Tanzim Hassan, who helped to write, test and tweak different clustering algorithms, Ziqiang Guan, who helped to set up a local geocoding server and implement final design features, and Carys Snyder who contributed initial sketches for the layout and design of the application.

The initial app pulls together data for five cities: New Haven, CT, Houston, TX, Raleigh, NC, Detroit, MI and Oakland, CA. Allan's team plan to share this early version of Clustr with these cities for feedback before expanding the functionality. Allan then plans to increase the data set to include many more cities.

The care that Allan's team has put into Clustr is clear — it really is remarkable easy to unearth patterns with Clustr and it's fun to use as well. We look forward to seeing how cities and citizens will use this data!