opendata sites: new in old style
From an opendata site we expect a great sort of datasets, as much as possible. We expect a great output of formats, a good geo accuracy, a data quality regularly updated/reviewed by the institution. An API to use remotely the data.
I also would like an standard open way for participation, a wikidata where people could improve, correct or add data in the same way as wikipedians do. The institution cannot improve alone all the data they have, is obvious that crowdedition data portals are needed (most great exemple is openstreet maps)
But what I really miss in an opendata site is a visualization of the data. Most of them they are gopher style, usually one level of categories and then a bunch of datasets inside each of the too general categories (data.gov,data.gov.uk, the datahub,…)
Bird view, please!
Before I search for something in a opendata site that I’m visiting for the first time, I would like to know some general information about what I have in front of me; I mean: please, tell me something about the data you collect and offer. Minimum important information is:
Terrace field in Yunnan, China, by Jialiang Gao
- number or records
- number of datasets
- average registers by datasets
- Some range, like min and max time covered in the registers
And then, I wanna see a picture of the whole thing. Can be an static image, can be a tool to filter data, to browse data, to find data; can be anything that helps the user understanding what are the site contents.
A lot of opendata portals are just online but not growing up too much. When transparency for politicians is just a branding topic, they open the opendata offices and that’s all; not too much budget and not enough opendata spirit.
Results: the prehistoric zone
Don’t you think that when you search for something, in a search engine, in a directory, in an opendata site, and you get a flat list of items with a non explicit reason for the order you see, is the most obsolete, prehistoric and non sense way to give you an output of resuts? Is not true that ajax and 2.0 concepts are everywhere except in query results? Isn’t that anachronistic?Retrieval Information field advances fast: better results, more efficient techniques, easier multi-archive search but the visualization of the data results is not advancing. In the specific field of opendata sites this is not different: I can search for a concrete geo dataset from my neighborhood, and maybe I find it, or I can ask for the dataset to the institution. This is ok if I come with a clear idea about what I’m looking for. I would like to be able to browse, filter, and see a representation of the datasets and then let the ideas come out.
I did some related works, like Area, trying to show the global view of a database and a simple browser tool.