Mapping Birmingham’s Food Scene

Our Approach

Starting with a list of about a dozen twitter accounts in Birmingham’s Food Scene, supplied by our friends @PopUpDosa, we identify who is most interested in the scene, then look at who else they follow in order to map the scene more fully.  This results in some very interesting maps with potential marketing implications for those involved.

The “seed” accounts were: @DigbethDiner, @themeatshack, @popupdosa, @smokeandumami, @ysl807, @SoulFoodProject, @brewdogbirm, @stirchleywines, @birminghamcubed, @TheWhiskyMiss, @yummy_brummie, @changekitchen

There’s a bit of a bias towards street food and savoury food, so I’d be interested to hear how I could extend the list, for example by adding @Suhaav.

Who’s Interested in Birmingham’s Food Scene?

Using our “seed list” of about a dozen food related accounts from Birmingham, we identified 133 individuals, businesses or organisations with a very high interest in Birmingham’s Food Scene. They range from the media side, such as Edible Brum Magazine (@EdibleBrum) and restaurant critic Paul Fulford (@PaulFulford), to baker and food evangelist Tom Baker (@loafonline), to blogger  and man about town (well, I see him at many events) Cy(ber) Birmingham (@cybrum), plus many more.

Click each image to see a larger version, or you can see all the details in this PDFs – which you can also search for names!

Map of twitter accounts interested in Birmingham's Food Scene

This map shows the accounts which follow at least 10 of the “seeds”. They represent people and organisations with a keen interest in Birmingham’s Food Scene.  Click for a bigger image, or you can see more detail in this PDF

The Unfiltered View

Next we look at who else these “highly interested” accounts follow…

This shows an unfiltered view of accounts followed by our "Highly Interested" group.

This shows an unfiltered view of accounts most likely to be followed by our “Highly Interested” group. Click for a bigger image or see full details in a pdf.

I’ve coloured the accounts using a clustering method, which uses information on who follows who to automatically divided them into three groups: red, green and blue. The red group at the top left contains many accounts which are generally related to Birmingham, but not so related to food. The other two groups are mainly food related, and you can spot features such as the cluster of hotels, pubs and city centre venues at the bottom left (in green) and accounts related to Stirchley at the top in Blue.

As we’re interested in the Food Scene, I’m going to cut out most of the red group. I’ll save a handful of “red” accounts which I know to be food related:

Four food related accounts hand picked from the "red" group. We'll keep these accounts in the analysis

Four food related accounts hand picked from the “red” group. We’ll keep these accounts in the analysis

Why are these accounts in the red group (and, for that matter, why are accounts like “Created in Birmingham” in the blue)? Some examples which might help explain: Dr Birgit Kehrer runs @changekitchen which has strong voluntary sector links, this has pulled the account over into the red group. Similarly, @edibleeastside , “An arts led initiative to transform a derelict site into a vibrant and contemporary urban garden…” has strong arts scene links. Note, that this doesn’t necessarily mean they don’t have stong food scene links, though this could be investigated separately.  Depending on how these accounts want to position themselves,  they may want to build more links with Birmingham’s Food Scene.

Another answer is that there is no perfect clustering, there are lots of ways to categorise accounts, and the automated process won’t necessarily agree with one individual’s view. However, I think you’ll agree it’s done pretty well!

Focussing in on the Food Scene

Dropping the majority of accounts in the “red” group, we re-organise and re-cluster the accounts specifically related to food. Let’s have a look. Click for a bigger image, or for a perfectly scalable one, which will look lovely when printed, download the  pdf.

Birmingham's Food Scene Mapped.

Birmingham’s Food Scene Mapped. Click for more details, or  see it all in a zoomable format (pdf).

So… what does this tell us about Birmingham’s Food Scene on Twitter?

Well, it looks like we’ve identified many of the key players. If we want to publicise something to the “scene” we’d do well to talk to those who are most followed (largest nodes). There are a number of other “Social Network Analysis metrics” which more accurately identify influence,  and when I run these, accounts such as Claire Tortise (@FredaPR_consult), a specialist food/drink communications consultancy, stand out more clearly.

There’s quite a high degree of geographical clustering, for example, Stirchley accounts can be seen in the top left with many Kings Heath nodes close by and Moseley nodes towards the bottom left. Extrapolating from the  Twitter evidence, we could contend that in Brum, many people are “hyper-locally” guided in their food choices, which is a fancy way of saying they want to eat somewhere close by.

Noticeably distance from other Moseley nodes  is Carters of Moseley  (@cartersmoseley) which is placed adjacent to another “fine dining” account, that of Glynn Purnell (@yummy_brummie).  Which seems sensible. We could interpret this as “people interested in fine dining are willing to travel”.

I’d be very interested in hearing your comments and interpretation of this map, either via twitter @AndyPryke (follow me for more similar investigations) or in the comments section below. I’m also open to suggestions for other sectors to map, and, of course, I’m available for commercial work in the area too (which comes with a lot more detail).

Reproduction Rights

You are welcome to reproduce these images as long as you reference either @AndyPryke or include a link to this blog post. I’d love to hear if you do. Mapping Birmingham’s Food Scene by Andy Pryke is licensed under a Creative Commons Attribution-NoDerivs 3.0 Unported License. Creative Commons License

Acknowlegements

  • R – Excellent language for data mining, analysis and visualisation.  Been using it since 2003 and it keeps getting better. Also the twitteR package by Jeff Gentry
  • Wordle – Producing the word clouds
  • doPDF / novaPDF – Great virtual PDF printer, used to “print” the Wordle output to PDF files.
  • XChange PDF Viewer – Very useful PDF viewer, allows annotations and easy to snapshot enlarged PDFs.
  • Irfanview – Lightweight image viewer / editor with enough features for almost all my image editing
  • Tony Hirst (@psychemedia) – The doyen of Twitter visualisation, whose tutorials have been very useful to me.

 

Advertisements
This entry was posted in Uncategorized. Bookmark the permalink.

15 Responses to Mapping Birmingham’s Food Scene

  1. andypryke says:

    I’d be interested to hear about clusters you’ve spotted, or nodes that seem to be in unusual places, and any theories as to why they’re there!

  2. Pingback: The Power of Networks and Network Visualisation: “Mapping Birmingham’s Food Scene” | SnookesPost

  3. Hiya Andy.

    This is a nice resource – especially for finding those that are good at cross marketing.

    The exceptions at the other end, however, are difficult to find. Can a subset this be generated on the fly from a given account (eg @loafonline) and is there any way of finding if a particular person is referenced (eg Neil Houston)?

    • andypryke says:

      Hi Mike, yes, you can generate a view from a single account (e.g. @loafonline), either by removing things not linked to @loafonline from the maps you see, or generating a new map just based on the account.

      To find if someone is referenced, download the PDF file(s) and use search within them. Note that some PDF viewers only look for whole words so “loafon” wouldn’t be found.

      Andy

      • Interactively? These all seem to be JPG or links to the relevant PDF…

        Some surprising omissions in there like Neil Houston, or even me…

      • andypryke says:

        The aim isn’t to identify those interested in food (except as a first step), instead it’s to list producers etc. People are listed if they are followed by lots of the people who were identified as highly interested in food, so you’re going to turn up more restaurants than individuals. Neil isn’t on there, but his @YetAnotherGin account is.

        You’d need the original gephi file to do interactive filtering, but unfortunately the version of gephi I use has a tendency to be unable to load files it saves! So the work would need re-doing first.

  4. dpoid says:

    Can you do this using force diagrams? Like I did here and here? (You might need Chrome or Firefox to render this properly.) Morcego is not in active development – and I’m not sure TikiWiki supportsit any more, but force diagrams are a lovely way to explore and write about connections.

    • andypryke says:

      Nice. I think I might have looked at these before. The layout of the graph is mainly done using the “Force Atlas 2” algorithm in Gephi, which is a very useful tool.

      • dpoid says:

        I like Gephi too, and tried to use it for some patterning/mapping around the recent urban food growing project activities. The wider link sets probably aren’t big enough to do the kind of aggregate thing you’re doing here.

      • andypryke says:

        Interesting! Is the work you did online? Are you on twitter?

      • dpoid says:

        The work I started on local food growing is very rudimentary, and I’ve set it aside because Kate Cooper is working on a related mapping project that would be much more powerful. I’m sure she’d be glad to hear from you. She’d be even happier if it meant a contributory piece on Growing Birmingham. Interested?

        As for Twitter, I havent got the presence of mind / patience to track it. My preferred social media is email, f2f and a couple of well-moderated talk boards. :)

      • andypryke says:

        Thanks for your comments – and the pointer towards Kate’s work. It’s reminded me of a whole sector Birmingham’s Food Scene which I didn’t target – sustainable food, allotments, grow based projects. I’ll have to include these in my next version of the map. Not sure I’ve time to write an article, but will contact Kate. Cheers!

  5. Elizabeth says:

    This is really good being useful, visually pleasing and easy to use for slow learner s like me! You should print some as posters for the biggies concerned to advertise. Good to see we made it!

  6. Pingback: Full to the brum – Birmingham Food Data Map

  7. Pingback: Twitter Map Portrait of TEDxBrum | The Data Mine

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s