Open data and diabetes

In December my daughter was diagnosed with Type 1 diabetes. It was a pretty rough time. Symptoms can start and escalate very quickly. Hyperglycaemia and ketoacidosis are no joke.

But luckily we have one of the best health services in the world. We’ve had amazing care, help and support. And, while we’re only 4 months into dealing with a life-long condition, we’re all doing well.

Diabetes sucks though.

I’m writing this post to reflect a little on the journey we’ve been on over the last few months from a professional rather than a personal perspective. Basically, the first weeks of becoming a diabetic or the parent of a diabetic, is a crash course in physiology, nutrition, and medical monitoring. You have to adapt to new routines for blood glucose monitoring, learn to give injections (and teach your child to do them), become good at book-keeping, plan for exercise, and remember to keep needles, lancets, monitors, emergency glucose and insulin with you at all times, whilst ensuring prescriptions are regularly filled.

Oh, and there’s a stupid amount of maths because you’ll need to start calculating how much carbohydrates are in all of your meals and inject accordingly. No meal unless you do your sums.

Good job we had that really great health service to support us (there’s data to prove it). And an amazing daughter who has taken it all in her stride.

Diabetics live a quantified life. Tightly regulating blood glucose levels means knowing exactly what you’re eating, and learning how your body reacts to different foods and levels of exercise. For example we’ve learnt the different ways that a regular school day versus school holidays effects my daughters metabolism. That we need to treat ahead for the hypoglycaemia that follows a few hours after some fun on the trampoline. And that certain foods (cereals, risotto) seem to affect insulin uptake.

So to manage the condition we need to know how many carbohydrates are in:

  • any pre-packaged food my daughter eats
  • any ingredients we use when cooking, so we can calculate a total portion size
  • in any snack or meal that we eat out

Food labeling is pretty good these days so the basic information is generally available. But its not always available on menus or in an easy to use format.

The book and app that diabetic teams recommend is called Carbs and Cals. I was a little horrified by it initially as its just a big picture book of different portion sizes of food. You’re encouraged to judge everything by eye or weight. It seemed imprecise to me but with hindsight its perfectly suited to those early stages of learning to live with diabetes. No hunting over packets to get the data you need: just look at a picture, a useful visualisation. Simple is best when you’re overwhelmed with so many other things.

Having tried calorie counting I wanted to try an app to more easily track foods and calculate recipes. My Fitness Pal, for example, is pretty easy to use and does bar-code scanning of many foods. There are others that are more directly targeted at diabetics.

The problem is that, as I’ve learnt from my calorie counting experiments, the data isn’t always accurate. Many apps fill their databases through crowd-sourcing. But recipes and portion sizes change continually. And people make mistakes when they enter data, or enter just the bits they’re interested in. Look-up any food on My Fitness Pal and you’ll find many duplicate entries. It makes me distrust the data because I’m concerned its not reliable. So for now we’re still reading packets.

Eating out is another adventure. There have been recent legislative changes to require restaurants to make more nutritional information available. If you search you may find information on a company website and can plan ahead. Sometimes its only available if you contact customer support. If you ask in a (chain) restaurant they may have it available in a ring-binder you can consult with the menu. This doesn’t make a great experience for anyone. Recently we’ve been told in a restaurant to just check online for the data (when we know it doesn’t exist), because they didn’t want to risk any liability by providing information directly. On another occasion we found that certain dishes – items from the childrens menu – weren’t included on the nutritional charts.

Basically, the information we want is:

  • often not available at all
  • available, but only if you know were to look or who to ask
  • potentially out of date, as it comes from non-authoritative sources
  • incomplete or inaccurate, even from the authoritative sources
  • not regularly updated
  • not in easy to use formats
  • available electronically, e.g. in an app, but without any clear provenance

The reality is that this type of nutritional and ingredient data is basically in the same state as government data was 6-7 years ago. It’s something that really needs to change.

Legislation can help encourage supermarkets and restaurants to make data available, but really its time for them to recognize that this is essential information for many people. All supermarkets, manufacturers and major chains will have this data already, there should be little effort required in making it public.

I’ve wondered whether this type of data ought to be considered as part of the UK National Information Infrastructure. It could be collected as part of the remit of the Food Standards Agency. Having a national source would help remove ambiguity around how data has been aggregated.

Whether you’re calorie or carb counting, open data can make an important difference. Its about giving people the information they need to live healthy lives.

Getting that learning fix

I’ve been doing some domain-modelling with an arts organisation recently. The domain model that we’re working on will help underpin a new version of their website.

We gathered some domain experts across the business and ran some workshops to start capturing how they think about their domain, what they do, what the outputs are, etc.

Data modelling isn’t something that most people do. The process of trying to build a model of the world is, to varying degrees, new to them. Just as understanding the nuances of different types of art events and curating exhibitions is new to me.

So there’s been an element of me teaching them about what I do — the geeky stuff — alongside them teaching me about what they do. By coming to more of a shared understanding we can collectively get more from the exercise.

I love that process.

I love learning stuff, particularly in domains that are very different to the one in which I often operate. You don’t often get chance to sit with a bunch of domain experts and hear them discuss what they do at great length.

I also love that light-bulb moment when you can see people suddenly get what you’re teaching them. Its like you can actually see them levelling up in front of your eyes.

(I’ve been trying to rationalise what, on the surface, seems to be too very divergent interests: a love of teaching & a love of coding; whatever the reason, it probably explains why I keep doing different roles).

I then got to thinking about how so many of the events that run these days are largely domain based: domain experts talking to each other. Not people teaching each other new things, maybe in wildly different domains. I guess Ignite events might be closest, but I’ve never been to one. They also seem like they’re highly structured and involve calls to action rather than teaching and knowledge sharing, but I might be wrong.

So what kind of events am I missing? Where can I get my learning fix, or is there scope for a new type of “geek” event?

Leaving Talis

Earlier today I hit the publish buttons on the blog posts announcing the shutdown of Kasabi and the end of Talis’s semantic web activities. Neither of those were easy to write.

My time at Talis — which will have been four years in September — has been a fantastic experience. I’ve worked with some incredibly talented people on a wide range of projects. The culture and outlook at Talis was like no other company I’ve worked for; it’s a real pleasure to have been part of that. I’ve learnt an massive amount in so many different areas.

I’d argue that Talis more than any other company has worked incredibly hard to promote and support work around the Semantic Web and Linked Data. And I’m really proud of that. Despite increasing — but still slow — adoption, the decision was made that there was only so much more that could be done, and that it was time for Talis to focus elsewhere. Over the next few weeks I’ll be winding up Talis Systems’ activities in that area, and working with existing customers on continuity plans.

This year has been very difficult, on a number of levels. On the whole I’m now glad that I can focus on the future with a fresh outlook.

In the short term I’m considering freelance opportunities. If you’re interested in talking about that, then please get in touch. My profile is on LinkedIn and I’m available for work from 1st August.

If you need help with a Linked Data or Open Data project or product, then get in touch. Over the past few years I’ve done everything from data processing through to modelling, product & technical strategy, and even training.

Longer term, I want to take some time to think about the kind of work that I enjoy doing. I love building products, particularly those that are heavily data-driven. I want to build something around Open Data. Beyond that I’m not yet sure.

If you have something that you think I could help with, then I’d love to hear from you.

Pastures New

Its been 9.5 years since I first started working for Ingenta. Over those years I’ve been presented with some fantastic opportunities and worked on some great projects with great people.
From a technical perspective I’ve developed a deep appreciation for hypertext, web architecture, XML, and semantic web technologies. I’ve spent the last 18 months or so creating a publishing platform that has semantic web technologies at its core. This is something I’m particularly proud of as we’re putting these technologies into production use. Our early experiences are that their flexibility is really going to pay-off when it comes to building next generation publishing and research tools. Meeting the changing requirements for researchers and scientists means really embracing semantic web concepts like linked, open data. So I’m confident that this platform is going to serve the company well in the future.
But I also decided that its time for me to move on and explore other opportunities. While my role has always been varied and changing, I decided it was time to do something different, in a role that would let me continue to work with semantic web technologies.
So I’m happy to say that from 1st September I’m going to be joining Talis as Programme Manager for the Talis Platform. I’ve been really impressed with what Talis have achieved over the past few years: they’ve got a real strategic vision and a hugely talented team. I’m excited to be joining them to work on developing the Talis Platform and help them deliver on their vision for the future. Its a natural step forward from what I’ve been working on for the last few years.
But first, time to relax with my family before getting my teeth into the new role. Exciting times ahead!

Five Things

The “five things” meme is still doing the rounds and it turns out I’ve been tagged by Phil Wilson. So here’s five things that you almost certainly don’t know about me:

  • My wife and I got together at University after a Rag 3-legged pub crawl. Yes, my charm is so bad that tieing myself to a lady with a skipping rope and plying her with beer for an evening is the best I can do. We’re still together 14 years though. Yes, I’ve untied her since
  • I once went to a hen night in Valverde del Camino in Huelva.
  • I have a degree in Biology and for my final year project I somehow ended up studying the egg laying behaviour of Callosobruchus maculatus. This involved sitting in a small, dark, hot room (35 degrees centigrade) which had no windows, watching six female beetles recording what they were doing every thirty seconds over a period of two hours. This was repeated 30 times. What I’d wanted to do was study gorilla behaviour.
  • I suffer from vertigo and am mildly claustrophobic (the latter not connected with the former, but it didn’t help much either)
  • The first ever database schema (OK, record format) I designed was on my first computer. It was an inventory system for my dad’s pigeon lofts

I’ll try tagging Geoff, Pete, Kirsty, Dorothea, and iand to spread things around a little further (and in some different directions!).