I was completely terrified. We were hiking downhill towards the Pacific Ocean, and I was fighting visions of slipping, tumbling down the hillside.
It was my first hike. In retrospect, it was not steep or challenging or particularly precipitous. But growing up in the flat plains of the Midwest, I had made it to my early twenties without a single hike (and was questioning the rationality of this strange subspecies known as “hikers”).
Many years later, I managed to hike the Inca Trail, where I mostly conquered my lingering fears of hiking.
Well, our data inventory was our Inca Trail, and let me just say…
Try some foothills before you scale a mountain
In other words - consider starting with something smaller than a full scale inventory (i.e. listing all of your data). In our case, our program had been around for awhile and the inventory was a natural progression. But for programs that are just starting, an inventory can take a lot of work without visible results. Ideally, you should have some successes or examples in place before starting an inventory.
Here are some ideas for alternatives to doing a full scale inventory:
- Topic Inventory. Collect and list all the data around a service or topic like Housing or Transit or Recreation and then pair it with a strategic release.
- Top Five. Ask departments to list their top 5 datasets. See DataLA.
- Department Dives. Dive deep into a handful of departments and inventory (and maybe publish) their data. See Philly.
- Initiative Inventory. Start with a specific problem or challenge your City is facing and inventory the data related to that. Again, couple with a strategic release.
- Stop at Systems. Simply identifying systems that generate data could be a natural stopping point. Our legislation was more specific and the dataset inventory is super helpful to have, but it’s one way to go.
Alternatively - do the full scale inventory, but use these alternatives as an implementation or phasing strategy. Whatever you do…
Break it into steps
Our legislative requirement to inventory our data didn’t come with a set of instructions. And we didn’t find a lot of guidance out there (though maybe we missed it). So we broke it into 3 phases:
- Identify data sources
- Brainstorm datasets
- Complete dataset inventory
Without these discrete tasks, the inventory felt too big and scary to accomplish. And for each of those steps…
Provide guidance and support
Our Data Coordinators (representatives from each department) ultimately had to deliver the inventory. (And FYI, they are the true heroes here.) So we did our best to support them every step of the way and make it as easy as possible for them to complete the work. During this phase your data coordinators are your users. Craft empathy for them and leverage user-centered design to support them.
Tools we provided:
- Guidebooks, webinars and workshops
- Shared definitions (e.g. data steward, data source) - thank you GovFresh
- Activities to motivate the process
- Templates designed to ease input
- Custom support and consulting for complicated departments
- A single online resource portal for Data Coordinators to go to
- Deadlines, including intermediate deadlines (they actually asked for this ;-)
Check out our Data Inventory Resources, but before you use them…
Implement quality control and maintenance
At each stage in the process, plan to allocate time and energy to ensure quality submissions. If you don’t do this, your inventory will be less useful. This takes a lot of work (try to minimize by thinking thoughtfully about your data collection strategy) but it’s worth it. Just budget the time. Also, come up with a plan for maintaining the inventory and always, always…
Track progress
We tracked the status of each department through each phase (and deadline) of the inventory. This forced us to checkin and helped us identify the need for targeted outreach and support. It also made us accountable to our legislative requirements.
At this point, 75% of our departments have completed the inventory, which you can check out here. We decided to go live with the inventory despite not having all departments ready. That’s also where tracking comes into play. We have rolled the tracking into our quarterly report and are accepting department submissions on a rolling, monthly basis.
Anticipate future needs: Inventory as platform
Your inventory is a one time chance to capture a lot of awesome information. Don’t waste it. Now that we have this tremendous resource in place, we are building all sorts of stuff on top of it. More on this in subsequent blog-posts but we are using the inventory to:
- Create publishing plans
- Automate monitoring and performance metrics in publishing and open data
- Conduct dataset management
- Build tools to endorse datasets for publication
- Provide data dating / data concierge services
- And more…stay tuned ;-)
And again, thanks to our amazing Data Coordinators for making this possible!