International civic tech startup DataLook is hosting #openimpact, a replication marathon aimed at helping API developers build local solutions with reusable, open source projects.
The project has so far attracted more than 150 developers from Europe, Asia and the U.S. to its Slack community channel, with four projects already replicated in new locations.
DataLook is focused on creating a directory of reusable data-driven projects centered on civic and social good. In the absence of API open standards for many civic projects, DataLook aims to shortcut the rhetoric and debate that can often surround government decision-making around what standards should look like by putting proven, effective solutions in the hands of new communities.
Now, to highlight the potential of reusable open source projects, DataLook is hosting the replication marathon, #openimpact, aimed at encouraging reuse of 10 key projects. Of these, six projects are built using APIs at their core.
Project: Food Inspections Forecasting
APIs used: Socrata Open Data API
How the project works: Predicts food establishments most at risk for the types of violations most likely to spread foodborne illness
Original city: Chicago
What it will take to be replicated in other cities: Where an API developer's city uses the Socrata open data portal (or another open data portal with API access to data sets), the Socrata API could be used to mine data in a similar way to how Chicago has made use of the data. A detailed blog post from DataLook explains how to incorporate data from a city’s API with R software to create prediction models.
How the project works: Allows citizens to report, view or discuss local problems
Original city: FixMyStreet was created in the U.K. by the not-for-profit mySociety and is used in locations around the world.
What it will take to be replicated in other cities: FixMyStreet has a sophisticated platform and set of resources that can help cities use APIs to integrate resident requests for problems to the local community to be reported and mapped.
How the project works: Searches tweets related to food poisoning and directs possible victims to a 311 service form
Original city: Chicago
What it will take to be replicated in other cities: A clear GitHub repo with sample YAML settings explains how it works in Chicago. Developers would need to add their Twitter and Google Analytics API keys and find a relevant Open311 API for their city, and they can then fairly easily replicate the project.
Project: DC Action for Children
How the project works: Maps child well-being based on publicly available data
Original city: Washington, D.C.
What it will take to be replicated in other cities: Developers can sign up for a Census API key and start creating maps using the census data and ranking/mapping terminology set out by DC Action for Children. Where a city has an API or data set with amenities like child care centers, libraries and healthcare, these could be mapped over the socioeconomic index in order to reveal whether there is an inverse ratio at work in the city planning system (that is, where higher-income neighborhoods are provided with more amenities and services than low-income neighborhoods).
Project: Cute Pets Twitter Bot
How the project works: Twitter bot for animal shelters that increases the likelihood that potential adopters find adoptable animals
Original city: Denver
What it will take to be replicated in other cities: A number of other cities have already deployed a similar API. Developers could use the adopt-a-pet API. By finding the URL for a photo library for an animal shelter in another city, the API could be adapted for use in other locations.
APIs used: Supports Open311 as a client and partly as a server API (see here and here)
How the project works: Allows citizens to subscribe to new legislation, to contextualize legislation and evaluate legislation
Original city: Philadelphia
What it will take to be replicated in other cities: A GitHub repo shows code for a beta API for linking legislation, council members and districts. Developers replicating this project would need to first find a similar set of API endpoints from their city. (This data may be available as endpoints in an Open311 API.)
So far, food security issues — in projects like Food Inspections Forecasting and Foodborne — seem to have the most traction with participating developers.
“We see that most people jumped on the Food Inspection Forecasting project,” says founder Tobias Pfaff. “My guess is that the project involves machine learning, and many people from our network are looking for projects where they can apply their machine learning skills. Apart from that, it’s a great project with proven impact. Helping the city to avoid food poisoning with your tech skills is just amazing.”
The project is part of a new platform approach that startups are considering when looking to build civic tech solutions. DataLook’s approach can also be seen in efforts like DataKind and Free Code Camp. All are focusing on linking developers and technologists with communities that need data and tech skills.
Pfaff says the #openimpact “replication marathon” is a test case that is finding an audience and credibility in the space:
So far we got some encouraging feedback from Chicago’s chief data officer, Tom Schenk Jr. We also got some support from a few local government officials in Germany that are dealing with open data. They tweeted about #openimpact and published a blog article.
However, Pfaff admits there is a lot to do in helping communities and residents recognize the potential of working with developers and startups to solve long-standing civic problems:
Since we launched #openimpact 10 weeks ago, we have not yet seen a direct collaboration between community organizations/civic leaders and developers. However, it’s super important to establish this connection so that the apps get used. Ryan Swanstrom stresses this in a DataKind blog article. He said that visualizations and open data are a "critical first step to bigger outcomes." We agree with how he said that "the strength of the partner organization that will ultimately use it to create change in the world."
Openimpact is a test run to see if we can scale this. We will evaluate in September what has worked and what hasn’t. One thing we saw is that there is a substantial crowd out there that is interested in working on replications. Some people immediately know what all this is about and start working on day one. Others take a look but then they are a bit lost, and I think we need to provide more guidance in the future. We saw people who were super engaged in reaching out to partner organizations and others who were just interested in solving the tech puzzle but didn’t invest time in looking for local partners to promote or use the app. Connecting developers and organizations is something we should maybe assist with in future replication marathons (or “replicathons,” as one user termed it).
DataLook’s #openimpact "replicathon" is open to contributions until Aug. 31. API Developers can join DataLook’s Slack community.