I am currently working on an open source solution(s) to publish open dataset(s) in a way a website could be quickly deployed to visualize and search this dataset(s)… I already made several websites serving several datasets with thoses solutions, and a oversight website is visible here :
With this project.s I tried to answer a couple of issues bothering me a lot relative to open data projects. Usually open datasets are not really understable / searchable by someone else than an engineer or a data scientist. To grasp an idea of how a dataset could be useful to the general public one could need a dedicated website (and a search enginne)… So I tried to create a CMS / something like a wordpress but for dataviz, offering the same structure to visualize various datasets with different kind of views we always encounter on websites : table, list, map, stats, … and the possibility to add static contents and customize the UI…
This project is still ongoing, and there are still a lot of emprovements I have to do, but as a POC I think it could be tested more… Let me know if you feel it could be interesting in your research
@julien_p It would be super interesting to split this topic into Middleware for Open Data & UX. You’ve done a fantastic job and the web site looks really nice. I’m curious about how you worked on UX to get to that point. I don’t have permission to split this topic into a new topic but an admin should be able to do that.
@paulineleon thanks a lot for your reply and advice. I’ll let an admin split the topic, but to answer quickly to your question about the UX I had the chance to work with an UX designer on a part of this project for 10 months.
At first we (the designer and I) were supposed to develop all this infrastructure only for a particular client (a non profit organization), all of that with public funding. This organization had a list of geolocalized items and they wanted a website to search within this dataset and a map… Anyway I insisted (a lot) to make all the open source tools we were developping more generic than they were supposed to be. So the UX designer had somehow (and I’m a bit sorry for her ) two clients instead of one : on one hand the non profit organization and the particular constraints of their own data, and on another hand me (!) insisting on scenarios like “what if they send us a completely different dataset ?”, “what if another organization want to reuse what we did ?”, “what if the organization or another wants to manage their dataset online ?”, “what if they want to change their logo or UI template by themselves, online ?”, etc…
Also we discussed a lot about the innterface simplicity we aimed for, to be as simple as for instance the interface of airbnb or instagram, despite having a lot a actions we wanted the user able to do (download a dataset, change the view, being free to implement the exact user’s UI, having favorites - that is not implemented yet, … ).
So it was quite challenging, but we agreed that open source tools that complex (map, filters, search engine, responsive, back office …) gains a lot to be thought as generic as possible from the start. Because if it’s open sourced, it’s supposed to be reused but not with the exact same constraints every time.