Lessons From London

The last week of June was a great week for open aid data. From June 28-30, Open Ag partners, software developers, donors and civil society representatives met in London (again!) to talk about tools for better data. Over the three day event, developers shared ideas and established connections between their respective data enhancement tools; users provided valuable insight for the design and functionality of such tools; and people from all over the data ecosystem networked, discussed, and mind melded on various topics related to the use of open data. At the same time, members of the International Aid Transparency Initiative’s (IATI) Technical Advisory Group (TAG) held a parallel meeting in London, colloquially known as a “miniTAG,” to discuss how to improve traceability in IATI data.

With so much to report out of our week in London, here are a few takeaways:

 

Visualization is key for data publishers too, not just data users. Currently, when publishing data to the IATI Standard, people don’t see what their data actually looks like before pushing it out to the world for all to see. Over the three day event, we heard that allowing people to visualize their data while in the process of publishing it, would help them improve data quality. It would not only allow them to see where they can improve, but also would get people excited about what they are contributing to the data ecosystem. It’s hard to find motivation to improve your data when you don’t know what your data looks like and, therefore, how it can be used. Publishing needs to be an interactive process for publishers to see its value, and thanks to some crack user research from Neon Tribe, we’re making sure to integrate this into our publication tools.

 

The Open Ag publishing toolkit will be perfect for mid-sized data publishers. Most large organizations already have the internal systems to publish higher quality data with a little guidance. Small publishers continue to rely on existing easy-to-use tools, such as AidStream, to publish data. But after talking with donors and civil society, the Open Ag toolkit in development appears to be the “Goldilocks” set of tools for publishing open data. For reference, mid-sized organizations are those for whom manual data entry would be fairly burdensome based on the amount of their activities, but whose data architecture is resource-constrained and therefore less sophisticated than bigger organizations. Our toolkit will include tools to validate your data in real time, tools to convert CSV data to IATI’s XML format, and tools that use machine learning to easily tag activities with sub-national locations and granular classifications. Prototypes will be ready by the end of July 2017 with a full launch in October 2017. All of our tools will be open-source.

 

Publishing data still causes heartburn–better tools can alleviate that pain. There is a lot of appetite to move away from harsh-sounding words like “accountability” or “IATI XML” when making the case for open data to general audiences.  Happier phrases invoking “knowledge and data management,” “mass collaboration,” and “smarter investment planning and targeting” are beginning to enter the lexicon; after all, people should know that there are concrete benefits for them to publish data, and we need to do a better job selling those benefits. However, many organizations still approach IATI with some trepidation, especially as it increasingly becomes a requirement from donors. This led us to a key insight: when you have heartburn, you don’t go the pharmacy for “feel good” medicine, you look for “heartburn relief.” We should absolutely remember that words matter when we talk about IATI to unfamiliar audiences, but many IATI publishers-to-be just want to know what to do next, who can help them, and how they can live up to their commitment to transparency and openness without undue burden.

 

There’s appetite for more. With so many interesting conversations, three days just wasn’t enough. Topics such as geocoding and the use of IATI organizational files (“hidden gems” says James Coe from Publish What You Fund) were of great interest. Moreover, there was much discussion on our new portal to access and visualize multiple open data sources. More meetings are likely to follow, but the Open Ag team was encouraged that so many people from all corners of the development and data world came together for this event. The inaugural IATI Mini-Tag was also a great success (check out these lessons learned  from DFID’s John Adams). Over the next few months, we’ll be building on this appetite as our tool development is in full swing. For those keeping score at home, our final set of tools to be developed is: 

 

●        An auto-geocoder that will improve location data by allowing data publishers to automatically tag projects with sub national locations.

●        An auto-classifier that will allow data publishers to easily tag projects with classifications (just crop types, for now) that are more granular than DAC sector codes.

●        A data validator that will assess a publisher's data and provide feedback on how to enhance the data to make it more useful.

●        A publishers’ toolkit that incorporates and stitches together all of the above tools.

●        A data portal that will allow users to quickly identify who is doing what, where and to what effect as well as compare IATI data to other data sets.

●        Enhancements to existing data tools, AidStream and d-portal.