It’s popular today to work in Open Data, Big Data or similar space. It feels just like the times of Web 2.0 mashups, but instead of simple Google Maps based tools we’re now creating powerful visualisations that often feel like an end by itself.
In a way, I expected OKCon, OpenKnowledge Foundation’s conference, to be about makers – people that build useful part of OpenData ecosystem and to provide in-depth case studies. Instead we’ve got a mix of representatives from government, large international NGOs, small NGOs and various developers and software providers. To me these worlds felt just too far apart.
Data portals are here and while we haven’t seen a large scale deployment from government in Slovenia, it’s likely that they we will have something in this regard in the next few years. Everyone else has already deployed their first version and more advanced (public) institutions are already on their second or third attempt. Just as with social media a few years ago, we’re now also seeing first case studies that show economical and political advantage of providing such sites.
Self hosted CKAN platform seems a popular tool for such efforts (or it just might be conference bias as it’s developed by OKFN).
Budgets and contracts
A lot of effort is expanded in area of representing budgets, tenders, company ownerships and similar. In this regard, Slovenia’s Supervizor looks like something that’s from far future compared to what other countries or projects achieved at the moment. We could contribute a lot back to the international community, if we can produce case studies on benefits (or lack of) of such system.
At the same time, building visualisations around local budgets is something that doesn’t feel productive anymore. I think we should just upload sanitised data to a portal like OpenSpending and focus our efforts on projects with more impact.
Just as with mashups, everyone loves geospatial representation. The more colours and points of interest, the better. The only problem is, that it’s often useless for people that actually need to use it. While not openly expressed during presentations, discussions during the break often revolved around how bloated and useless were these representations and that just having a good text/table based report would work so much better.
At some point, community will have to embrace modern product development methodology – stories, user testing, iterative development and similar. Right now it feels like a lot of these tools are either too generic or sub-contracted and developed through water-fall model.
Having said that, I’ve seen great examples of how to do things right: landmatrix.org
While NGOs might be building things the old fashioned way, their developers certainly aren’t. Tools and platforms are openly licensed and published on GitHub and often tied into different continuous integration environments.
- https://github.com/ostap/comp – automatically exposes your local CSV, XML and similar files as JSON endpoint through standalone Go based server. Developed by mingle.io team.
- Drake – for building workflows around data
- Pandas – python based data analysis
Software development is already hard for teams of seasoned veterans that work on projects inside the tech industry. It’s almost impossibly hard for both large and small NGOs since there just isn’t enough talent available. Additionally, it seems that these organisations often don’t want to coordinate efforts (even basic sharing of data) with each other or even internally, making projects even less likely to succeed.
I think that we’ll continue to see a lot of badly executed projects in this area until modern, tech-driven groups like OKFN and Sunlight Foundation manage to raise the bar.