No coding required?

18 October 2008 (Saturday)

Reading a thread on the mysociety list, which discusses a posting showing a mashup technique “with no coding required” and a subsequent critique of the specific mapping presentation to which this technique was put, led me to think about how screen scraping adds or removes value from it source data. My first response was that the two parts of the online discussion – technique and usage – were separate evaluations that shouldn’t be conflated.

After all, doesn’t the technique just guarantee an easier point of entry to the practice of screen-scraping (which mysociety’s sites themselves rely on)? A technique “with no coding required”, “an easy way for people to get into this sort of thing if they don’t have the coding skills of others”.

But. the two questions – technique and usage – are related, because if no shining example of the technique can be found, is it a valid technique? Perhaps one might argue, some people should be focused on technique (this is part of the domain of, say, software development), whereas others (statisticians, information designers, authors) should be focused on usage… and thus a technique might be neutral.

While described as “no coding required”, a glance at the original post, shows that, while no specific programming language is being used, there is coding going on (see Bonnie Nardi’s description of excel as a user-programming language in “A Small Matter of Programming”). The coding that’s going on is a form of re-presentation – taking data from one format (a textual table) and translating it onto a map.

The usage of similar forms of screen-scraping by mysociety websites is in the form of taking multiple sources of data and rather than just representing each, finding new value by aggregating (and/or re-partitioning them) in novel ways. This requires all the trappings of typical “coding” – hard work development involving persistent databases, coping with long-term changes in data and managing servers and server-loaded code.

All screen-scraping inevitably involves some stripping out of data (even if it is only of original context). There must be someway of adding back some new value for any screen scraping to

And it is precisely this “coding” that enables the value in mysociety sites. Is it possible that the technique, which doesn’t add value in the form of aggregation or repartitioning (filtering) but just “re-presents” the data, might be used in a context where radical value was added simply through a new presentation form, but I wonder if that’s possible… with ‘no coding required’.


2 Responses to “No coding required?”

  1. Tony Hirst Says:

    I’ve been thinking about this too – and have just posted a ‘re-presentation’ of data that is being made available via an API (and as such is accessible ‘only to developers’) in a spreadsheet (which although still requiring users to have some IT skills are more familiar to more people):

    The data is US campaign finance data from the New York Times Campaign Finance API ( )

    (Note that I didn’t check the T&C and of the API to see if this was an illegitimate use.)

    The mysociety services are a great example of how to take data from one or more services, adding value, insight (maybe through partioning or segmenting) and/or providing additional context along the way, and then making it available to end users in a familiar form. (I subscribe to a feed of my MP’s Parliamentary questions, for example).

    But as more and services open up data via APIs, and wonder whether we can (and should) also be making the data available via “People’s APIs”, such as spreadsheets?

  2. freesteel Says:

    This issue relates to why we deleted the attendance rate column from the table of MPs on the publicwhip site. We had seen one too many mash-ups based on this particularly meaningless field of data, and wanted it to stop. People weren’t getting the idea that they should do something which related to what the votes were about. Why is this so hard?

Comments are closed.

%d bloggers like this: