What happens when you use the AI image engine DALL-E for a blog post?
I was just finishing a post on Medium titled "APIs — the best way to train AI on transport & mobility" and needed an image to exemplify this.
So I purchased a few credits on DALL-E as an experiment and gave it the prompt:
"a computer scientist on a train holding a bag containing data"
(Yes, I did use the word 'train' there in the transport context, rather than the education one - in an attempt to be slightly funny)
And this was what is came back with:
I then picked the second image for the post, as I thought it represented most closely what I wanted... even though the rendered face was slightly incorrect:
The concept of Open Data has been around for many years (the earliest I can find is around 2000, but there may well be much earlier usage). Yet some individuals still seem to be confused by the concept.
Here is the Open Data Handbook’s definition:
"Open data is data that can be freely used, re-used and redistributed by anyone - subject only, at most, to the requirement to attribute and sharealike."
https://opendatahandbook.org/guide/en/what-is-open-data/
This clear and understood approach to data provides increased transparency, improved decision making, fosters innovation and can increase civic engagement.
But its range of use can perhaps make it difficult for people to understand what Open Data is.
Here’s some examples:
Research:
Researchers & academics can use Open Data to study impacts and changes across economies, governments and populations
Business:
Organisations can Open Data to develop new products and services, such as a map app that shows the location and availability of parking spaces
Public Sector:
Governments can use Open Data to demonstrate and improve transparency & accountability, such as by publishing data on Public Sector spending.
Individuals:
Citizens and journalists can use Open Data to hold governments & authorities accountable, such as tracking the performance of local schools.
However, in my experience some people have deliberately misunderstood the term Open Data. Why?
Open Data can lead to change, and some people may be afraid of this. They may worry that Open Data will subsequently make it more difficult for them to control information, perhaps leading to job losses. Others may have vested interests in keeping data closed. For example, businesses may not want their competitors to have access to their data, and governments may not want citizens to have access to data that could be used to hold them accountable.
Ultimately Open Data is a powerful concept. It is therefore up to all of us to ensure that we all share the same definition and push for Open Data's continued societal benefit, not just the few.
The rise and rise of Generative Artificial Intelligence (AI) has been the main technology talking point of the last 6 months. Text services like ChatGPT and Google Bard are now able to engage in dialogue that may evenpass the Turing Test. And image generation services such as DALL-E & Stable Diffusion can create realistic and almost life-like pictures just from a text description.
To perform these feats means training these large language
models (LLMs) on vast amounts of data that is computationally expensive
and therefore prohibitive for many organizations.
For example, OpenAI's GPT-3 LLM was trained on a dataset of
175 billion words. This dataset was collected from a variety of sources,
including books, articles, websites, and code repositories. Google's PaLM LLM
was trained on a dataset of 540 billion words,
But as the flurry of initial excitement about Generative AI now
dies down a bit, focus is turning to the sources of data used. With the worry
being that unless that data is reputable and trustworthy, these systems will have
the wrong inputs to base their machine learning algorithms upon.
Perhaps more than ever, the computing phrase of “Garbage in,
Garbage out” is worth more than a passing consideration.
Mobility-as-a-Service (MaaS) is more than just a multi-modal journey planner and ticketing app. Integrated digital transport platforms have the power to provide seamless & easier mobility options and incentivise users towards more sustainable means of transport. I personally believe that (interoperable and open standards-based) MaaS is the future of transport for those living in cities, towns and more rural communities.
I'm therefore incredibly happy that I have been asked to join the Steering Board of MaaS Scotland. To provide external strategic guidance towards the group's ambition to support the implementation and scaling-up of MaaS solutions across Scotland.
In an earlier blog post I explained some advantages and disadvantages of the "Plan Buy Run" [PBR] model / approach to IT management.
Note: Yes, I know the term is usually "Plan Build Run" - but how many larger organisations these days build everything from scratch
This approach is more "waterfall" in its delivery, which can be effective for organisations that need a clear and structured approach to IT. But it also means they are more subject to the typical issues of:
PDB and agile can run in a hybrid way across the Plan, Buy and also the Run phases.
For example:
The plan-build-run (PBR) IT model is a traditional approach to IT management that divides the IT function into three distinct phases:
The PBR model has several advantages, including:
However, the PBR model also has some disadvantages, including:
The PBR model is therefore seen as a waterfall approach to IT management, as each phase is completed sequentially before the next phase begins.
I've not been using Blogspot for about a year, as I have been using other blogging platforms instead.
Why?
Well I've noticed that this blog is getting less & less traffic and also less traction in search engines (with some posts falling out of Google entirely).
Perhaps its best to just use this platform more my informal thoughts before posted articles that get more traffic & engagement on Medium, LinkedIn, etc.
Some people argue that social media sites, such as Facebook and Google, should pay for the news content that they use, as it helps to support journalism and the production of high-quality news. Others believe that such a move could be harmful to the free flow of information on the internet and could lead to less diversity in the news. Ultimately, the decision on whether or not to make social media sites pay for news would need to be made by individual governments based on the specific circumstances and considerations of their own countries.
However The New Zealand government has now announced that it will introduce a law mandating that big online digital companies like Alphabet’s Google and Meta Platforms pay New Zealand media companies for the local news content that appears on their feeds.
https://thenewamerican.com/new-zealand-mulls-over-law-to-make-facebook-and-google-pay-for-news/
Its time to confess all folks!
The little secret that most Search Engine Optimisation (SEO) agencies still don't want to acknowledge is that SEO still works even when you stop employing the agency... but with one huge caveat.
So in the same way that organic rankings improve over time from effective optimisation efforts, sites tend to take time to drop lower and lower when SEO is ignored.
How quickly these rankings drop depends considerably on the site, the competition and a number of other factors. And my experience has been that continual sustained SEO efforts have better success that turning that optimisation effort on and off over time.
*Unless they have used paid inbound links and stop these straight away. In which case, you shouldn't have hired that agency for SEO in the first place!
Yesterday I posted (for the first time in ages) that I had found a useful diagram called the "European Interoperability Framework" covers different levels of interoperability.
This work from the European Commission in March 2017 gives specific guidance on how to set up interoperable digital public services. https://ec.europa.eu/isa2/eif_en
It documents 47 clear recommendations on how to improve governance of interoperability activities and ultimately improve the quality of European public services. And in my opinion this includes the transport & mobility sector.
The "European Interoperability Framework" diagram clearly makes the distinction between four layers of interoperability governance: legal, organisational, semantic and technical
Legal interoperability:
This is from legislation that suits not only the physical but also the digital world. Its purpose is to facilitate interoperability between public services at the lower levels
Organisational interoperability:
This provides consistency of processes for optimised data exchange
Semantic interoperability:
This ensures that the precise format and meaning of the data and information exchanged is preserved and understood between each party.
In other words that data entities & elements and the relationship between them are clear.
Technical interoperability:
Ensures that systems are linked to each other correctly for the exchange of data. This ultimately requires the use of formal technical specifications.
I not only run a digital strategy & marketing consultancy (Ideal Interface), but also chair a transport data standards organisation (The Open Transport Initiative).
Unsurprisingly I do speak a lot about Interoperability and its benefits. However I recently found this very useful diagram in a document from the University of Ghent (or Gent, if you speak Dutch):
https://biblio.ugent.be/publication/8504205/file/8504206
The diagram, annotated as the "European Interoperability Framework" covers different levels of interoperability.
Marketing efforts to get new customers cost money, either in the form of resource budgets or media budgets (or both). This metric, known as the Customer acquisition cost [CAC] is an important commercial metric that calculates the total sales & marketing effort used by the number of visitors that convert or buy a product / service. This figure, along with Customer Lifetime Value (LTV), are probably the two most effective and comparable marketing metrics used.
Understandably most digital marketing managers also like this overall average cost to be broken down by each acquisition channel. This allows them to assess the effectiveness of each, with the usual groupings given by digital analytics applications including:
Organic search can have the lowest CAC
Typically Organic Search still delivers a lower CAC than other marketing channels, such as paid search or display advertising.
This has been explained in detail in many posts, e.g.
https://www.growthcollective.com/blog/customer-acquisition-cost-ads-seo
https://firstpagesage.com/seo-blog/seo-roi/average-customer-acquisition-cost-cac-by-industry-b2b-edition-fc/
Although my own word of caution when calculating CAC is that all costs must be counted (e.g. media, agency management, in-house staff coordination, etc.) not just some, if you are to properly compare "apples with apples".
In a recent post, I described what a Data Marketplace is (e.g. an online portal or website where data can be easily found and shared)
Since then I have been working with a company to create "a consolidated market place for multiple data sources specifically targeted for the transportation and mobility market"
https://www.smartpassenger.co.uk/
I therefore thought it would be useful to map out the capabilities a data marketplace would have.
These capabilities are generic descriptions of key features you could expect to see in any Data Marketplace, regardless of the sector or industry vertical the platform was found.
A data marketplace (or data mart) is an online portal or website where data can be easily found and shared.
Although the name indicates that data is only bought and sold via a Data Marketplace, other data can also be made available such as Open Data (free to use, re-use or redistribute) or Shared Data (limited by license or end-usage).
Data marketplaces typically offer various types of data for different markets and from different sources. Making them a focal point for providing and accessing information about specific sectors or industries.
There can be a wide range of data formats and technologies provided via Data Marketplaces. However the widespread adoption of APIs for integration has meant that these are often the chosen modern way of publishing data sources now.
Lately I have been looking at data quality in the transport & mobility sector.
Helpfully UK Gov has 6 "Core data quality dimensions":
1. Completeness
describes the degree to which records are present
2. Uniqueness
describes the degree to which there is no duplication in records
3. Consistency
describes the degree to which values in a data set do not contradict other values representing the same entity
4. Timeliness
describes the degree to which the data is an accurate reflection of the period that they represent (and that the data and its values are up to date)
5. Validity
describes the degree to which the data is in the range and format expected
6. Accuracy
describes the degree to which data matches reality
These terms are use across UK Gov and were created with the UK Data Management Association https://www.dama-uk.org/
More info here: https://www.gov.uk/government/publications/the-government-data-quality-framework/the-government-data-quality-framework
I'm finding that I'm increasing using the military term VUCA when referring to a technology project. This acronym, which stands for volatility, uncertainty, complexity and ambiguity has been used over the last 2 decades to reflect the lack of knowledge about what is about to happen.
For me it sums up the vauge definition of large software-based projects, where either there has been insufficient strategic input, not enough analysis or just a general lack of willingness to define future scenarios or risks.
I'm now on the hunt for further useful military terms, without falling back on SNAFU.
We live in a business world now driven by Digital Transformation. It seems that every media outlet wants to be "the next Netflix" and every takeaway restaurant thinks they should be "the next Deliveroo" Note: This is despite Deliveroo actually losing over £200million last year, a fact that hasn't been lost on the UK stock market in the week
2020 will be remembered as the pandemic being the catalyst for a huge shift to digital. Digital working processes, Digital meetings, Digital retailing, etc.
It will also be remembered as the year there was a huge reduction in the overall use of transport, as:
- those working were strongly encouraged to work from home, where possible
- lockdowns prevented a significant amount of personal & business travel
- tourism was stopped
- people were furloughed from work
Photo credit : https://unsplash.com/@thenightstxlker
But, as vaccines are rolled-out and restrictions are lifted, we are likely to see a significant "bounce-back" in the transport & mobility sector over the next few months, as staff return to the city and the office (although perhaps not to pre-pandemic levels for a long while / if ever).
So now it is high time for transport operators (and their suppliers) to fully embrace digital transformation and modernize their legacy business processes & systems. The current situation of low passenger volumes may mean reduced revenue, but also means less risk when making significant technology and data changes.
In short, if you are going to implement a new process or technology solution across the transport sector... now is probably the best time to do it.
I see it in nearly every new website set of requirements these days… the request for “personalisation of the digital user experience” or something similar.
(Usually just sitting in one line of a very long set of
crafted and prioritised requirements or user stories – hiding in wait to catch-out
the unprepared software or digital agency that has to fully respond with a
costed and carefully caveated proposal in a matter of days)
So what is meant by “personalisation”?
In my experience there are two different types of digital personalisation:
Implicit personalisation
This is where the user experience is changed based upon inferred
and non-personal details cleaned from the user e.g. their referral site, the search
engine term they used, the language of their device, the day and time of their visit
and even the assumed location they are browsing from.
This gleaned information can then be used to serve-up more
tailored marketing messaging & content (text & images) and perhaps more
targeted products that have been previously viewed or purchased by similar
people. The aim is that the website ‘learns’ what assets to serve to improve the
site’s goals (AKA the conversions), typically an ecommerce purchase or a lead
generation form completion.
It can be relatively easy to implement basic implicit
personalisation, either using such functionality already available with a decent
Content Management System (CMS) vendor or from a Conversion Rate Optimisation tool
that can be added subsequently.
Explicit personalisation
This functionality has the same basic aim, to increase the
conversion rate of the digital experience and therefore make more money or
generally get more business. This is also done by serving up the content, data or
features to favourably change the user’s behaviour. However explicit
personalisation does this by utilising data actually known about the customer,
such as: their address data (for more local content), their demographic data
(for more age or gender relevant content) or their previous browsing & purchase
data (for more targeted content).
This form or personalisation, as you can imagine, requires
the digital user experience to have access to some or all of the customers' personal data, perhaps stored in an online account. It therefore typically needs a more complex integration to the source of the customer’s data and secure
handing of potentially personal details.
