Re: Online #KnowledgeManagement - audit and suggestions for change #knowledgemanagement


Dear all-

I had a very nice talk today with Oleg Lavrovsky, who makes his living as an independent open data consultant ( He organizes most of the open data hackathons in Switzerland, and is active in sustainability data as well, so I think his feedback is worth listening to. Here are the points I noted:

We desperately need a getting-started guide - this should be the first thing one sees, after the small motivation section. We want to communicate from the beginning that each visitor has something to contribute, and that we need their contribution.

As a part of this getting-started guide, it would be worth building a sort of toolbox for beginners to bonsai, like the one found here: The idea is that we will have a bunch of people looking at our site who aren't experts in coding or semantic tech, but who are motivated to participate. If they see a wall of complications, they will just close the tab and move on, but if they have a list of tools with good docs and tutorials that will give them a chance to start contributing in a meaningful way.

In his experience, hackathons work best if there is one focus. In our case, I think it is worth re-emphasizing that the goal of the hackathon is to have a complete working demo that we can show people, and it probably makes sense to focus on getting the existing work packages to be good instead of starting a bunch of the stretch goals (though I have been dreaming of a bunch of BONSAI components:

He felt that the hackathon was a good opportunity to already involve more people online, and that there would be an audience if we did more outreach. Online participants need to be able to work independently, which means that we should have some example small projects that people can just claim and work on.

As a follow-up to this, I think it would very good to have a list of projects where we need help - this would allow people to either pick a project, or to get a sense for the kind of projects BONSAI works on, so that they can define their own. Here are some examples I can think of:

  • Simple API that would take barcodes and a region, and spit out an LCA score and contribution analysis.
  • Guess labels from BONSAI product and industry names based on a machine learning classifier. Keeps original data, just adds new columns. Takes CSV or RDF as input. Training data will come from the descriptions and documentation of the ontologies.
  • Extract the car model from (maybe combining with GREET: and make it web-accessible and following BONSAI coding standards.
  • A set of global and regional independent validation data for LCA/MFA, e.g. how much CO2 was produced in 2018, how much land was transformed, how much shipping happened, etc.

A key to hackathon success is to make sure that everyone can make meaningful contributions. Each person should be a champion or leader of their task/deliverable. Key here is fostering a positive and welcoming atmosphere, and being flexible as groups evolve.

One idea that was new to me was to have a data-driven hackathon, where we would be measuring data about the hackathon itself. Oleg and his team have developed, and though it is difficult to explain, it is quite cool to see in action. He would be willing to set up an instance for our hackathon.

Finally, we only get one chance to make a first impression, so all our communications should reflect and emphasize our core values, especially our embrace of openness, both in consuming open data, and in making our data available for others. We want people like Oleg to see our website and feel at home, knowing that this is a place where he could make a difference. It isn't easy, which is why we need to start ASAP.

I see that Oleg has joined our list, so he is welcome to add/edit the above points :)

Join to automatically receive all group messages.