Jump to content

Jason Mackinlay

Registered
  • Content count

    5
  • Joined

  • Last visited

Community Reputation

0 Neutral

About Jason Mackinlay

  • Rank
    Participant

Personal Information

  • Organisation Membership
    EA

Recent Profile Visitors

98 profile views
  1. Scaleable small data insights

    One of my colleagues, Hugh McCann, recently presented our company with an update on his data analytics work. He kicked off with a summary of his home energy usage. I knew he'd been tracking this, and I believe we talked on the day he accidentally deleted 3 months of data! Yep, he's a nerd's nerd and hadn't thought to protect the data because it was just a home project, so you can imagine how he felt. He's just published a more detailed description of how and why he did it here. What immediately strikes me is that what he's done isn't particularly complex and it appears to me that it could be commoditised / built into new constructions. This ties in with some other discussions on smart metering but gives some insight into what can be done across a broader range of domestic power usage, and a way to engage the home owner. It seems an obvious extension to add a degree of machine learning and mass energy usage collection for a whole variety of purposes.
  2. Scaleable small data insights

    One of my colleagues, Hugh McCann, recently presented our company with an update on his data analytics work. He kicked off with a summary of his home energy usage. I knew he'd been tracking this, and I believe we talked on the day he accidentally deleted 3 months of data! Yep, he's a nerd's nerd and hadn't thought to protect the data because it was just a home project, so you can imagine how he felt. He's just published a more detailed description of how and why he did it here. What immediately strikes me is that what he's done isn't particularly complex and it appears to me that it could be commoditised / built into new constructions. This ties in with some other discussions on smart metering but gives some insight into what can be done across a broader range of domestic power usage, and a way to engage the home owner. It seems an obvious extension to add a degree of machine learning and mass energy usage collection for a whole variety of purposes.
  3. Pirelli's smart tires - what's next?

    I agree Tim - the first generation phase in IOT should be a gimme for most applications. And not particularly interesting or informative. I think the challenge we face is that some of the things we want to use IOT data streams for are actually emergent properties. And we often need to be collecting the gimme data products from several elements in a system before we can start to interpret the bigger picture. For tyres, I can see feedback mechanisms for fuel efficiency and suspension wear based on what an instrumented tyre is reporting. And if you apply this to the transport industry, you could look at predictive indicators for tyre failure as simple measure, or tie it into the Mass Management system for increased payloads. Link all of this to an IOT enabled road and some active tyre / vehicle control mechanisms and you could have big improvements for safety, capacity and maintenance costs. But I won't know if I can do any of those things until I start measuring the components to see what causal relationship exists between each sub-system. So on this, I'm not concerned that Pirelli are just selling tyres, as that's about all they can do with what has been established. And this application is just one 'thing' in the IOT.
  4. What does it take to be an IoT engineer?

    @Andrew at MEA you've made a great point here and it's easy to see how your approach differs from that of other companies. Most of my work relates to using data rather than generating it and I place a value on good data (specifically data with integrity, currency, completeness, correctness, quality). One common classification of the data associated with a task is to consider the data products in three ways: the raw data collected or generated, the analysis and deliverables given to the client, and the knowledge generated from doing the work. It is good practice to describe in the contract how each of these products are treated for ownership. It is a reasonable approach for a company to sell data as a service, and this is a different business model to what you describe as the MEA approach to sell hardware that collects data. I disagree that you are facing an ethical dilemma about whether to sell data or not - producing a data set of temperatures does not infer ownership of the weather. I do agree that farmers could greatly benefit from wide area sharing of this type of data with their neighbours, and I would like to think there is a viable business model for those companies that could incorporate data sharing. The mutual benefit would outweigh any perceived commercial advantage to exclusion. I'm sure these systems already take advantage of open source data which are often provided under Creative Commons licences (see example at BOM page). This provides a simple solution that goes back to data integrity - it would make an exciting disruptor to this environment if 'hardware' companies such as MEA incorporated data sharing under a CC arrangement as an inherent characteristic.
  5. Data Analytics

    Big data is often confused with lots of data. This topic raises the issue of how to handle the growing volumes of data, and the 'Flood of data' knowledge note shows just how large this can be, but it there is further discussion that can be had on the differences between handling high volume data as a technical issue, and conducting data analytics of a big data set. The link can also be drawn to IOT design whereby the purpose for sensing and observing a particular data event should be a deliberate outcome of the design process to inform a decision. I see reports of data stores being overwhelmed by the volume of data collection and without a plan to reduce or analyse the key elements.
×