There’s plenty of talk about data, analytics, Big Data, artificial intelligence, deep learning and so on. Nerdy conversations that tend to keep the geeks, the marketing department and the press release writers happy but the rest of the population completely cold.
Who’s the Real Data Audience?
Let’s remind ourselves where the rest of the population actually are.
Can you guess? A hint for you, they reside with two standard deviations of the average and make up the majority of people.
All the talk of open data, developers taking to social media to give the likes of Translink asking for their god given right to open data, all very well but it doesn’t resonate with the key stakeholders…. the public, the businesses and the day to day humdrum of Northern Ireland.
There’s excellent work going on with the open data initiatives from DETI and other interested parties. Progress may be slow but I’d expect it to be slow (with expected public service cuts don’t expect Translink data to be high on anyone’s list). I know the tech heads are itchy to do things and hackathons are happening (the Urban Hackathon is coming up at the end of June). The real questions are these: Does it resonate with the public? Where’s the win win? What’s in it for them?
With all the talk of opening data up, “do you do open data, everyone does open data”, there’s little talk of the potential data collaboration between small medium enterprises (SME’s) in the province. Does is matter? Yes of course it does.
A Simple Collaborative Example
Let’s take a hotel, Hastings Hotels operate a number of locations in the province. Is it possible to predict room rates 40 days out based on certain factors? Of course it is, there’s last years bookings and repeat visitors. That’s looking back, I’m more interested in predicting forward. Assuming occupancy is at 80% what will it take to hit 90%?
Now I could rest on laurels and assume that Game of Thrones is going to push up the numbers, stick my finger in the air and see which way the wind is blowing.
Even better though would be to take a feed from somewhere that has plenty of rich event data, large scale events and smaller ones the area. With a feed of dates, event types you could calculate the peak nights of occupancy. Data from What’s On NI (http://www.whatsonni.com) is about as rich as it gets, local events, big events and major events get listed. That data has value.
So the question is: Taking a feed from whatsonni.com can I (Hastings) calculate room rates for the next 40 days based on peak event data?
I believe it’s possible and a win for both parties, whatsonni.com could gain revenue from the feed for each of the hotels and if Hastings could raise peak pricing even by 7-10% on an average time of year room rate, the multiples involved would be a big win for them.
Concluding
I’ve dreamt up one example, a simple but highly effective one. It’s an easy sell to both parties, “I’d love both of you to increase revenue by collaborating, let’s do a trial for six months”. Now have a think about all the other businesses out there, data interconnecting and collaborating with each other. A series of paid for end points where everyone else could potentially benefit. This sort of thinking will raise NI’s bottom line and it’s all possible.
It’s also a perfect fit for proof of concept grants, where there is a solid basis of potential to see real benefit in all business sectors, not just development of what I would consider limited use mobile applications.
You’ll still need the help from the nerds, there’ll always be a need.
It's All A Bet - MLOps, DevOps, AI and Synthetic Data Design 
2 responses to “Harvesting Data Collaboration in Northern Ireland #data #startups #prediction”
Interesting, esp the collaborative PoC example, however I’d be conscious of the difference between “how many people does it assist” versus “how much money can biz make out of it”. The Data Science is very good at the latter almost automatically, but Open Data is the limiting factor on the former. Having known and worked with several people who’ve been burned by our local “MINE” mentality when it comes to semi-public operational data, it generates not only a feeling of disenchantment and pointlessness, but also perpetuates and environment that stifles any form of experimentation.
Well, that’s my procrastination done for a while… Keep up the awesomeness!
Interesting example Jason. I can see the problems and difficulties in, 1. expecting open data to come quickly and to give fast returns, and, 2. find a balance between open and profitable. Perhaps there needs to be a bit more networking, coordination and genuine commitment locally to achieve mutually beneficial returns from collabs such as your own here.
It also depends on what the difference is between the data and the product. Is the data the product? Then it’s unlikely to be made available to others. But if data can be used to enhance existing or future services, you’ve got a solid case for collaboration. The rest is down to businesses to concentrate on how they can improve their end products.