Feedback

[spjika]

Worthy idea.

1. I'm not convinced it should be done at a state level given these services are typically not under any state jurisdiction - no direct responsibility means hard to mandate activities. The state could build a framework and support it though. A better role for the state would be to financially support a local agency/org to manage this service.
2. Only minor duplication issues, but yes.
3. Crowdsourcing is partly sufficient - more important though is to gather bulk data from local orgs who have collated/maintained it and help them compile/dedupe that data first. Crowdsourcing is especially useful in long term validation and updating of data though.
4. I doubt this is reasonable no. It currently only happens when an agency is financially reimbursed for this task- hard to see this scale as a freebie. This can be a serious workload- just vetting changes isn't enough to keep things right, there needs to be some pro-active effort to update and vet info- if someone closes or changes services, if no-one bothers to submit a change, the data stays bad. Need a financial incentive to ensure this happens- but crowdsourcing content makes this less expensive than current 211 models use. Win.
5. It would help overall efforts- as long as y'all are connecting to those efforts, all good.
6. It could create a mess yeah- in every community there is some local org/agency who collects some of this stuff- they need to be involved. So this scales slowly... A local community would need technical assistance to gather and clean current data, would need help refining a taxonomy for local use, language adjustments, and would then need some support to use this new system- web devs to spin up portals etc, and training on data cleaning and vetting.

more thoughts later.. this is hopefully food for thought rather than just dumb criticism. There is a case for something here!

[roughani]

@spjika these are all solid points! Thanks for sharing.

[greggish]

Generally agreed with Spike's points.

1) Roles I see for the state:

• Publishing data on services that are provided by and funded by the state. There are a few directories already in the CHHS portal; surely there are others.
• Aggregating directory datasets that are themselves aggregated by state agencies and intermediaries (like 2-1-1s) that receive state funding.
• Developing tools and protocols that ensure the ongoing accuracy of the above.
• Providing various kinds of support to help underfunded institutions shift to open/standardized platforms.

2) A federated model is, I think, precisely what's needed in the long-run. The challenge is building that road. Ideally, state and local agencies and NGO aggregators can collaborate in the short term to take incremental-but-valuable steps that demonstrate the concepts, while learning and deliberating about what kinds of federation are needed to make this more and more effective.

3) I think it's worth getting specific about the 'crowd' here. I believe the most likely members of this particular crowd here are going to be service providers (librarians, social workers, healthcare practitioners, etc) and database administrators of organizations that aggregate this information. If you can develop a system that is trustworthy enough to be valuable for them (presumably one that is designed for interoperability so that they don't have to abandon their existing system to benefit from a new thing) and can receive their feedback, then the incentive will be that they can get better quality data by sharing that update alongside all the other service providers and directory aggregators than they would if they were just on their own.

4) There are probably multiple people in each county who already do this work -- alone in siloes. 2-1-1s alongside organizations that produce subdomain-specific directories (legal services, etc). What kind of system would promote stewardship that's shared among multiple institutional stakeholders?

5) I can say that the key objective of the Open Referral initiative in this next phase is the development tools and interfaces that enable distributed stakeholders to collaborate in the production of reliable data. So, it could very much help if properly scoped!

To interrogate the principles listed in your next steps a bit: while I don't think that conforming to a particular data standard should be an a priori objective of this kind of project, I also don't think that it's as simple as 'making the information available as quickly as possible.' Important information might not be getting collected or datized in the first place, and simply focusing on 'getting the data out' may miss the step of assessing what data should be out there. Likewise, if there are old directory datasets out there, containing information about services that are already being better maintained by NGOs like 2-1-1, then 'getting the data out there' could make a mess. If your goal is to work towards a world in which this data is accessible, interoperable, and trustworthy, then you may find that adopting a standard format is a necessary (though certainly not sufficient) step along the way.

[roughani]

Thanks @greggish! Super-helpful feedback.