Paul Wilton

Knowledge Graph

How knowledge graph technology is helping Cochrane respond to COVID-19

Julian Everett

Over the past five years it has been our privilege to be a technology partner of 
Cochrane [https://www.cochrane.org/]. For anyone unfamiliar with their work,
they are globally recognised as the gold standard in evidence-based healthcare.
Historically their primary published output has been systematic reviews, which
evaluate and summarise the current state of human knowledge about specific
clinical questions. These then inform both healthcare policy and clinical
guidelines for point-of-care decision making around the world. For example, over
90% of WHO guidelines are informed by Cochrane reviews. 

The review production process starts with the formalisation of the clinical
question of interest in terms of the Population it relates to (e.g. age, sex,
condition), the Interventions being investigated (e.g. drugs and/or procedures),
the Comparators against which the Interventions are being evaluated (e.g.
placebo), and the Outcomes against which effects are measured. This is then
translated into search protocol and run against trial registries and journal
databases to identify all relevant clinical study information in the public
domain. The matches are then evaluated for risk of bias, and the subset of
results which meet Cochrane’s stringent data quality thresholds are subsequently
aggregated and analysed to determine the cumulative effect estimates. 

In essence, this is about the identification of causal relationships between
interventions and outcomes. To that end, up until now Cochrane has focussed on
double-blind randomised controlled trials. They provide the highest quality
input data: randomisation greatly reduces the likelihood of alternative
explanations for the observed results as the number of participants increases;
blinding of field researchers and participants (i.e. hiding information about
specific interventions) removes the cognitive and behavioural bias risks that
might arise if, for example, placebos are known. 

Finally, review production is not a static process. New information is being
released into the public domain all the time both in the form of additional
journal articles about completed studies as well as information about new
studies. As a result, ongoing evidence surveillance and timely meta-analysis
updates are important additional requirements.

During our time working with Cochrane, we have built a new data platform using
domain-driven design techniques which has an explicit information model encoded
into a knowledge graph at its core. The model is published here
[https://data.cochrane.org/ontologies/]. This is complemented by a loosely
coupled, service-based architecture which ensures each business capability in
the value stream is cohesively encapsulated. In a follow up article, my
colleague Paul Wilton will review how the technical architecture hangs together
[https://datalanguage.com/blog/how-knowledge-graph-technology-is-helping-cochrane-respond-to-covid-19]
. The platform design allows services to be very quickly deployed in new
choreographies as the business context changes - something which has proved
invaluable over the last few weeks. 

So, what have been the business function and information flow impacts of the
Covid-19 pandemic on Cochrane? Firstly it has created an unprecedented critical
demand for the most up-to-date, reliable evidence to inform frontline clinical
decision-making around the world in order to save as many lives as possible. To
that end, they have been asked by the WHO to produce a precisely scoped,
authoritative register of Covid-19 studies. This essentially inverts their
business model from one based on a wide coverage of clinical questions using
only the highest quality primary evidence to one based on a single clinical
question which urgently extracts every bit of informational value from every
piece of evidence available. That means using all types of RCTs and
interventional studies, as well as observational studies, modelling studies,
diagnostic/prognostic studies and anything else that might improve outcomes. The
surveillance requirement for updates and new evidence switched from months to
days. The cycle time for systematic review production needed to be cut, and a
new Rapid Review format has been developed as a result to ensure baseline
datasets can be pulled together as quickly as possible and then augmented with
streamlined editorial content to meet clinical need. At the same time, the
feedback loop to frontline workers was closed to ensure that their highest
priority clinical questions are the ones being addressed first.

The Covid-19 Study Register [https://covid-19.cochrane.org/] was launched
internally on Monday 30th March and publicly on Monday 6th April. In all, two
back-end platforms were integrated together and a new front-end web application
was designed, built and shipped in just under 3 weeks. It has necessarily been
our fastest ever delivery by some distance, and was only even remotely possible
because of three core enablers:

 1. A data architecture, technical architecture and continuous integration
    infrastructure which minimised the technical risks and impacts of adapting
    to the changing organisational context.
 2. A small number of individually, amazingly talented developers.
 3. The tacit/emergent team structure and pre-existing relationships between
    those developers, established over a long track record of successful
    delivery on other demanding client projects, which meant that as a delivery
    unit they functioned at a truly optimal level from day one.

Since internal launch, our focus has been on getting as much data as possible
into the information model by any and every means we can think of, as quickly as
possible and at the highest possible quality levels. This is covering expert
crowd annotation, machine learning, and other text analysis and informational
reverse engineering techniques. One of the biggest automation/ML challenges
currently is the lack of training data due to the rapid emergence and rate of
evolution of the Covid-19 domain. We have previous experience of this class of
problem from our work with media newsrooms and automated classification of
breaking news stories. It is likely the crisis will drive extensive further
innovation in this space. 

At the same time, we have started receiving comments from stakeholders and end
users across the organisation. It has been some of the most inspiring and
touching feedback we have ever received both as an organisation and as
individuals, especially in light of the wider situation right now. The high
point so far was one very senior Cochrane staff member who, in wonderfully
memorable terms, said he couldn't overstate how useful the register is to all
aspects of his work and that he's now going to shout about it to everyone who
will listen.

A beautiful moment in the midst of demanding times.

Technical Strategy

Covid-19: Rapid Response Technical Delivery at Cochrane

Steve Liles

React and SVG make a powerful combination: Declarative UI component library
meets declarative graphics language — a match made in front-end developer
heaven.

In this mini-series we’ll be looking at rendering interactive graphical UI’s
using components built with React and SVG.

These days SVG icons are ubiquitous, but those are not the sort of components
we’re talking about.

We’re going to put together the interactive floor-plan UI using simple React
code that renders SVG elements instead of HTML. We’ll take it relatively slowly,
building through the basics into more advanced topics.

 * Part 1 — Declarative Graphics (this post)
   In which we look at the match made in heaven that is React and SVG
 * Part 2 — Arcs, Angles, and Transformations
   Keeping geometry simple with groups and transformations
 * Part 3 — Clip-paths
   An adventure in composite geometry
 * Part 4 — Masks
   When clip-paths don’t cut it: masks to the rescue
 * Part 5 — Animating components 
   Where we animate individual components
 * Part 6 — Animating the viewBox 
   In which we fluidly zoom and pan our entire UI

Let’s dive straight in!

Declarative Graphics
React developers enjoy first-class support for HTML elements in JSX:

const SomeComponent = () => (  
  <div>
    <p>hi!</p>
  </div>
);

But did you know that this applies to inline SVG elements too?

const SomeOtherComponent = () => (
  <svg viewBox="0 0 100 100">
    <circle cx={50} cy={50} r={30} fill="red"/>
  </svg>
);

Out of the box we can write  components that compose SVG primitives to create
complex interactive UI’s, just as you would with more typical HTML based
components.

(Note for React-Native folks: you’ll need to install react-native-svg, which was
looking pretty good last time I worked with it. Here
[https://stackoverflow.com/a/39041997/473338] is a simple but pretty powerful
example I created a few years ago).

Simple composition of declarative graphical components (https://62zqd.csb.app/)
Notice that we did all this drawing without a single line of imperative code,
and the declarative markup of our ghosts is very readable and accessible to
anyone familiar with React — a far cry from what this would look like if we
tried to do it with the Canvas API.

All your elements are belong to us
To draw anything with SVG, we first need an <svg>element to define the drawing
context within which all descendant elements are rendered.

We can have many <svg> elements in our page, each describing an isolated drawing
context — e.g. one <svg> per icon, chart, or other snazzy graphical widget.

Alternatively, we can root many graphical components within the same <svg> and
create an entire UI in a single drawing context.

Viewport and ViewBox
At a minimum an <svg> element should specify how much screen-space it will
occupy — its viewport. This can be done with CSS styles or width and height
attributes.

<svg width="200px" height="200px">...</svg>

We’ll typically want to specify a viewBox too, to explicitly define the
“user-space” portion of our drawing that is made visible in the viewport.

Think of the viewport as a window into the world defined by the SVG
(“user-space”), and the viewBox as zoom and pan settings that control what parts
of the user-space you can see through the window.

The viewBox attribute is a string which specifies, in order, the user-space x
and y coordinates of the top-left visible corner, and the width and height of
the viewBox.

<svg viewBox="x y width height">...</svg>

For example, “-100 -100 200 200” describes a square viewBox 200 units high and
200 units wide, centred around the origin at (0,0).

<svg width="10" height="10" viewBox="-100 -100 200 200">...</svg>

If we don’t specify it explicitly we get a viewBox that maps width and height
1:1 to the viewport, with the user-space origin (0,0) at the top-left corner.

To re-iterate: the viewport describes how much space the <svg> element takes up
within your HTML page. The viewBox describes the portion of the image that is
visible within that viewport. By changing the viewBox we can zoom in or out, or
show entirely different portions of the SVG image.

The only difference between these two SVG’s is the viewBox and the background
colour (https://0oy6p.csb.app/)If the proportions of your SVG viewport in pixels
don’t match the proportions of the viewBox, the default behaviour is to preserve
the aspect ratio and centre and fit the viewBox within the available viewport.
This behaviour can be controlled by varying the preserveAspectRatio
[https://developer.mozilla.org/en-US/docs/Web/SVG/Attribute/preserveAspectRatio] 
attribute.

By default the viewBox will be centred and “fit”within the SVG viewport (
https://52fbc.csb.app/)The separation of coordinate systems between the viewBox
and the viewport means that within our SVG we can work with coordinates that
makes sense for our specific drawing.

For example, if we’re drawing an interactive widget we can make the maths easy
for ourselves by working in percentages and specifying a viewBox of “0 0 100
100”.

Or if our widget has radial symmetry and we want to easily draw from the centre,
we can use a viewBox of, say, “-50 -50 100 100” centred at the origin (0,0).

In this series we’ll be drawing floor-plans, so we could choose to work in
metres, or feet-and-inches, or whatever system of measurement our plans use.
We’re actually going to use millimetres, just for the simplicity of working in
whole numbers.

We’ll describe our floor-plan with some simple JSON, specifying the room
geometry as coordinates of corners working clockwise around the room.

Now we can set up our viewport and viewBox, allowing a little padding around the
floor-plan so it isn’t too snug in the viewport.

import data from './floorplan-data.json';
const App = () => (
  <svg 
    width="500px"
    height="500px"
    viewBox="-1000 -1000 14000 11000"
    style={{backgroundColor:'blue'}}
  >
    // ... floorplan components here ...  
  </svg>
);

With our viewport and viewBox sorted, we can get on with drawing.

Declarative Drawing Primitives
SVG includes a number of drawing primitives ranging from the simple <line> and 
<circle> to the more flexible <path>. Mozilla Developer Network has a good 
reference [https://developer.mozilla.org/en-US/docs/Web/SVG/Element] of the
available elements.

Shapes can be filled and/or stroked, where fill is the colour applied within the
bounds defined by the shape, and stroke is the colour applied to the outline of
the shape. The width and other styles of the stroke can be controlled with
attributes or CSS.

We’ll be drawing our initial floor-plans using two simple primitives: <line> and 
<circle>.

<line> describes a line between a pair of (x, y) coordinates.

<line x1={100} y1={100} x2={200} y2={200} stroke="red"/>

<circle> describes a circle of radius r centred at (cx, cy).

<circle cx={50} cy={50} r={20} fill="green" stroke="black"/>

Room with a View
To draw our building we’ll pass the data into a Floorplan component which will
render out the various elements of the floor-plan — initially just the rooms.

const Floorplan = ({ data: { rooms } }) => (
  rooms.map(r => <Room {...r} />)
);

For the Room component to draw the walls we’ll need to map over pairs of
consecutive corner coordinates, linking them together to form walls. For example
a rectangular room with corners a, b, c, and d, will have four walls
(a-b),(b-c), (c-d), (d-a).

We can pull out those pairs with a simple function, perhaps wrapped in a useMemo 
hook for efficient re-renders.

const walls = useMemo(() =>  
  coords.map((_, i) => {
    const a = coords[i];
    const b = coords[(i + 1) % coords.length];
    return [a, b];
  }),
  [coords]
);

Now drawing the walls is as simple as using the SVG <line> element to describe a
line between the pairs of connected corners.

That works, but it isn’t very pretty. Let’s extract a Wall component and a 
Corner component, and get a bit more creative.

In our wall component we’re now drawing two lines per wall — first a thick white
line, then over-drawing the white line with a thinner, dark blue, dashed line.

Corner is very simple — just a circle, stroked white and filled dark blue.

The Room component now simply composes these higher-level Wall and Corner 
components instead of SVG primitives.

https://codesandbox.io/s/react-svg-basics-rk1t3You may have noticed that we
nested our Walls and Corners in an SVG <g> element. This is a logical grouping
element — the <g> on its own doesn’t render anything visible — but it does give
us a neat way of performing transformations such as rotations and translations
on entire sub-hierarchies of our SVG, and a means of capturing events from those
sub-hierarchies by attaching event handlers to the <g> element.

Recap
In this post we saw that:

 1. We can easily draw arbitrary shapes using SVG primitives and
    straight-forward, declarative React code — no plugins, libraries, or
    imperative code required.
 2. Complex graphical UI’s can be built by composing together increasingly
    higher level components, just like any other React application.
 3. The separation of viewBox and viewport lets you draw in a coordinate space
    that suits you, then render the result in a viewport appropriate to the
    final display requirements.

In “Part 2 — Arcs, Angles, and Transformations”, we’ll dig further into those 
<g> elements and look at paths and transformations while adding camera’s to our
floor-plan.

Until next time, happy drawing!

Software Development

Graphical UI's with SVG and React,  part 1 - Declarative Graphics

5 latin characters
Until recently, my understanding was that apart from 5 latin characters GraphQL
has got little or nothing to do with Graph Databases. But things are changing,
Graph Database vendors have started bringing these two technologies together. A
recent project has given me some insights.

For some background, GraphQL is a contemporary API interface standard used as an
alternative to REST. GraphQL was originally developed by Facebook. It provides a
structured query language for querying (and mutating) data, along with a runtime
implementation to invoke those queries upon your data. In pretty much all
implementations it uses JSON as its serialization format.

REST (I suspect most readers know this) is a lightweight low-level API standard
for performing uniform and a predefined set of stateless operations on data
aligned with the HTTP verbs. With REST you have to design and  build your API
interface contract for that set of operations - and while of course there are
common patterns in REST APIs, this is very much arbitrary. REST makes no
assumptions about the high level API contract or the underlying data model. The
developer defines both. With GraphQL you only define the data model (the GraphQL
schema), exposing to the GraphQL runtime and your consumers what the shape of
the data is.

With a GraphQL query you can request only the minimum data required to fulfil
your current need. Of course you can do this with a REST API too - in fact it is
a common pattern with JSON REST APIs to filter response data by JSON fields
using dot notation. But of course you need to build this yourself.

GraphQL does a lot more for you to boot, as well as out of the box query
patterns for mutating data, it also supports subscriptions (pub/sub style) - so
your GraphQL client can subscribe to server side events, and have data pushed to
it.

The developer ecosystem around GraphQL is superb now - The Apollo GraphQL
[https://www.apollographql.com] implementation works beautifully with React -
integrating with React Hooks for query and subscriptions. Apollo also has
in-browser developer tools akin to the redux-devtools for GraphQL debugging, and
you can manage local app state via Apollo using the GraphQL schemas / queries
etc.

Caching and performance
Requests to GraphQL are typically all POST requests - not great for caching you
ask ? Indeed, this is where REST has excelled for years. But the GraphQL server
implementations have come to the rescue. The Apollo GraphQL server has an
internal caching layer that supports full response caching, and fine grained
field-level caching (read more here
[https://www.apollographql.com/docs/apollo-server/performance/caching/] )

Amazon’s AWS AppSync is a fully managed GraphQL service built upon the Apollo
technologies - and can source data out of the box from other services such as
DynamoDB, ElasticSearch, REST APIs etc. With fully managed caching and real-time
data subscriptions, it really does make your developer life easy.

On the surface it still appears GraphQL has little to do with Graph Databases.
Yet increasingly we are hearing about how they should go together like bread and
butter, sausage and mash.

Not all graph databases are equal
I won't go into depth about what a Graph Database is, but am going to explore
some of the options available, and the relationship between Graph Databases and
GraphQL.

Graph databases come in several flavour combinations. First up you have RDF
graphs vs Labelled-Property graphs. The first big choice. Both are very
different [https://en.wikipedia.org/wiki/Graph_database#Graph_models]. Your
primary use case is important here. Essentially if you primarily need to walk,
query and analyse paths through your graph, then Labelled-Property graphs are
probably the way to go. If you are more interested in the nature of the
relationships between things, and creating rich semantic representations of
things, then RDF graphs hit the spot. Both types have their own well-developed
query language. Labelled-Property graphs are typically queried with Cypher or
Gremlin (eg Neo4J), RDF graphs are queried with SPARQL.

While most mainstream RDF based graph databases are fully compliant to the RDF
and SPARQL W3C standards, not all RDF graph databases are equal. Each has its
own implementation nuances and come with their own added extras (decisions,
decisions!):

 * An RDF graph database may or may not support reasoning (to some degree of the
   RDFS and OWL standards).
 * Reasoning may be at query time (backward chaining) or at write-time
   (forward-chaining).
 * The way reasoning is implemented can dramatically impact performance.
 * Add-ons and extensions for GeoSPARQL, Path analysis, Full text search,
   connections to other datastores etc.

To give some examples (DISCLAIMER - apols to those vendors not mentioned, I am
only referencing those I have worked with properly hands-on, and this is not an
exhaustive features list - it’s just a flavour) :

GraphDB [https://www.ontotext.com/products/graphdb/] (Ontotext)

 * Fully compliant with RDF/SPARQL standard;
 * Highly compliant and configurable OWL / RDFS forward-chaining reasoning
   (materialises inferred knowledge at write time);
 * Unparallelled full-text-search capability using lucene, or via bi-drectional
   connectors to external SOLR or ElasticSearch clusters.
 *  Ability to materialise post-inference fully structured JSON or JSON-LD
   objects into ElasticSearch indices or MongoDB
 * Geospatial querying and other extensions..
 * Rich visual query and management workbench UI

Neptune [https://aws.amazon.com/neptune/] (Amazon AWS)

 * Fully compliant with RDF/SPARQL standards;
 * No reasoning capability
 * Fully managed cloud service
 * Highly, highly horizontally scalable (primarily due to lack of reasoning)
 * Full text search capability via connector to ElasticSearch - materialisation
   into elastic is constrained (you cannot materialise user-modelled JSON
   objects into an elastic index), but delivers good full-text search utility.
 * You use this either as a Labelled-Property graph (using Gremlin) or an RDF
   graph (with SPARQL)
 * No query workbench/admin UI - you need to install your own 

Stardog [https://stardog.com]

 * Fully compliant with RDF/SPARQL standards;
 * Highly compliant and configurable OWL / RDFS backward-chaining reasoning
   (materialises inferred knowledge at query time);
 * Extremely fast online write times (due to backward-chaining model)
 * Adequate full-text-search capability using internal indexing
 * An outstanding SPARQL extension for PATH analysis (Pathfinder) - you can
   query for paths, shortest paths with weighted edges, normally only possible
   with labelled-property graphs
 * Geospatial querying and other extensions..
 * Rich visual query and management workbench UI

So you can see that each of these three are fully W3C compliant RDF / SPARQL
graph databases. But each has one or more killer features, that the others do
not. 

GraphDB - the post-inference connectors to ElasticSearch and MongoDB unlock some
outstanding technical and data architecture patterns out of the box.

Stardog - combines the semantics of RDF with sophisticated Path querying and
analysis.

Neptune - fully managed cloud service with unparalleled horizontal scalability,
but no reasoning.

The upshot - choose wisely based on your requirements now, and what is likely to
matter most in the medium term.

..and then enter stage right, GraphQL

Sausage and Mash or Pineapple on Pizza ?
How do GraphQL and Graph Databases fit together? As we learned earlier, GraphQL
is a REST API alternative with a structured query language combined with a
runtime for query processing and serving data to GraphQL clients (and it is
almost always implemented with JSON in mind).

So why are Graph Database vendors getting involved and what do you get for your
money?

The cynical marketer in me says those 5 latin characters definitely count for
something. Which sales manager wouldn’t love the potential energy pressure in
the ready-to-explode singularity of Graph Databases and GraphQL (while humming
‘Don’t stop me now’).

However, GraphQL does only provide another query mechanism for your graph
database alongside SPARQL or Gremlin.  Moreover it is much less expressive, and
most definitely not designed for exploiting the rich semantics of RDF and OWL. 

In fact if you go all in on a michelin-starred Graph Database, and only use
GraphQL to query it, I would argue you really probably don’t need a Graph
Database - a NOSQL JSON store would probably suffice. Of course, if you have
forward-chaining inference in your graph database (as with GraphDB), then your
GraphQL queries will benefit from the RDFS/OWL semantics you have modelled your
data with, which does add weight to the value-proposition.

Still hungry ?
Ontologies are an important part of the picture. In the RDF world an ontology
semantically describes the model and nature of your domain, and ultimately the
data structures stored in your knowledge graphs, and if inference-aware, they
describe the meaning of the data, and how we reason upon it. Typically these
models are communicated up and down our tech stack (and the responsible
developers), and often all the way out to consumers of your APIs (public or
enterprise-internal).  While we typically move data around using JSON the
linked-data aficionados will know well that if we do this using JSON-LD, we can
use the same ontology models from the knowledge graph (or a subset) for
describing the JSON-LD payloads we ship around outside of our graph database.

So do we need GraphQL to take advantage of these ontological graph models ?.  No
we don’t, we can do it with normal REST APIs (querying our graph with SPARQL)
and using JSON-LD payloads - but of course we have to build more API code to do
this.

With GraphQL though we can (we must) formally describe our data structures as
GraphQL schemas, and so we have a great opportunity here to align our GraphQL
schemas with those same ontology models that describe the data in our knowledge
graph -  this could be either a subset of our ontology landscape, or possibly
the superset, if we are using other GraphQL (served) data sources aside from the
data in the graph database. This is starting to look like very nice decoupled
data architecture with well structured cohesive data models, with possibly the
most efficient mechanism for consuming data in our web/mobile applications (a
GraphQL client).

It is looking more tasty
Adding GraphQL client interfaces to a Graph Database does make life easier - if
you are rapidly prototyping straight from the graph, or federating data into an
external GraphQL server there is a nice pattern for making a common set of well
described data models.

The Graph Database vendors have indeed jumped on-board the GraphQL sushi train
(and it’s not all about those 5 latin characters). Stardog and GraphDB (from
Ontotext) both offer GraphQL clients now.

At this point I was going to show an example of how GraphQL might be used to
query a Graph Database, but Jem Rayfield at Ontotext has already done the job in
his superb Star Wars blog post
[https://www.ontotext.com/blog/the-rise-of-the-knowledge-graph/] (if you have
got this far, then this is a must read).

Ontotext has taken the next logical step and created a ‘Semantic Object Service’
 that automatically generates GraphQL schemas bound to the underlying RDF data
model  in GraphDB, moving us towards the ontology ubiquity I described above,
where your domain ontologies are expressed in the knowledge graph, and across
your wider data-estate, including your GraphQL schemas. Nice work.

Jem concludes by saying :

> GraphQL is a very popular graph query language, well used, known and liked by
developers. The Ontotext Platform uses GraphQL to lower the barrier of entry to
knowledge graph data, whilst still providing the full expressivity and power of
SPARQL.
and this is a key message, which equally applies to Stardog and any other graph
database that may yet offer a GraphQL client. While GraphQL lowers the barrier
to entry, GraphQL gives you only as much as it is able to with respect to
querying and graph traversal;  to really unlock the power of the graph database,
you will need to use SPARQL.

But if you never plan to, and you are not interested in reasoning either, then
possibly you should ask yourself if you really need a graph database ?

What if you want to use GraphQL as the primary API technology in your stack, but
want to use SPARQL too to access the graph database (amongst other tech)?

A mixed data technology case study
In a very recent engagement, the Data Language team have been working on a Smart
Cities IoT project. The nature of the domain meant the type of data we needed to
deal with was very ‘graphy’ in nature, and the ability to describe our data
semantically and traverse a graph with SPARQL was of high utility. Inference was
less of a concern. The proof of concept for the project was brought together in
AWS with :

 * AWS MSK - a high performance Kafka message bus to consume, orchestrate, and
   distribute IoT sensor events
 * AWS Neptune to give us our scalable RDF based smart city knowledge graph
 * A set of JSON-LD REST API endpoints invoking SPARQL queries on the graph
   running in Kubernetes (EKS)
 * A suite of Kafka event processor consumers and producers deployed into
   Kubernetes mutating the graph, and emitting new events based on our business
   logic rules.
 * AWS AppSync - the GraphQL engine sourcing events and the JSON-LD REST APIs
   from Neptune, providing the primary developer interface to our client
   applications
 * A ReactJS built UI using GraphQL react-hooks to subscribe to real time events
   data, and to query data from the sensors and our knowledge graph (ultimately
   via expressive SPARQL queries).

A simplified view of the architecture looks essentially like this :


The kicker was modeling data across the entire stack using a suite of
ontologies. These ontology models describe our knowledge graph data, our IoT
message payloads in JSON-LD and were all federated in the primary application
API expressed as GraphQL schemas in AppSync. Building the ReactJS front end
using the Apollo GraphQL client and react-hooks was a very nice developer
experience.

Dessert
While GraphQL technically has little to do with Graph Databases, they are not a
perfect match like sausages and mash, GraphQL is just another API / query
technology, and not as expressive as the native graph database query languages
like SPARQL, but in itself is a high-utlity, very sweet technology.

There are some clear synergies, and do provide some of that warm comfort food
feeling when used together. GraphQL does give you an easily understood and very
developer friendly way into a graph database. Once you add the secret sauce of
domain ontologies describing both your knowledge graph and possibly even your
entire data-estate, expressed through your GraphQL schemas, then you have a very
nice pattern for a decoupled data architecture with well structured, nicely
communicated data models across your technology stack, and out to your consumers
too - making for efficient application development using the GraphQL ecosystem.
Get stuck in.

Graph databases

GraphQL and Graph Databases

Fabio Colasanti

Earlier this week, this tweet
[https://mobile.twitter.com/random_walker/status/1196870349574623232] appeared
on my feed.

> Much of what’s being sold as "AI" today is snake oil. It does not and cannot
work. In a talk at MIT yesterday, I described why this happening, how we can
recognize flawed AI claims, and push back. Here are my annotated slides: 
https://t.co/iCpyFw5urN pic.twitter.com/pmOTI3vq8p [https://t.co/pmOTI3vq8p]

— Arvind Narayanan (@random_walker) November 19, 2019
[https://twitter.com/random_walker/status/1196870349574623232?ref_src=twsrc%5Etfw]

Prof Narayanan's deck
[https://www.cs.princeton.edu/~arvindn/talks/MIT-STS-AI-snakeoil.pdf] has
reached a relatively sizeable audience and received widespread positive
commentary. After all, anyone who has had any involvement with AI in the last
few years can 100% relate with the catchy title of his deck.
Hey - we're just about starting to see a feeble light at the end of the
blockchain tunnel, right?

But, but...
Something doesn't quite sit well with me. So I'm going to stick my neck out and
respectfully reject, if not the conclusions, the way Prof Narayanan has chosen
to make his point. Or rather, the way he has chosen to make his real point.

I find this deck deceptively coy, as the actual message is not "How to recognize
AI snake oil" but rather "Why I don't want AI anywhere near socially sensitive
topics", which would have been a perfectly valid argument, and made for a
fantastic set of slides, but required a slightly different set of arguments and,
most importantly, to be explicit about it.

At the end of the day ML/AI is about identifying patterns in data and draw
inferences, so of course it can "predict the future". Google Maps does it all
the times when it gives you an ETA for your journeys. In Prof Narayanan's own
examples, bail decisions and border control are not performed by machines, but
based on the prediction of machines. It's not about predicting the future: it's
called decision support, performed at speed that is beyond human capabilities.
Is the way banks and insurance companies make decision on mortgages and premiums
snake oil too? 
Prof Narayanan seems exceedingly preoccupied about "massive transfer of power
from [...] workers to unaccountable tech companies" and points at "enormous
commercial interests" that "we must resist" and "push back". He is of course
completely entitled to give the whole argument a political spin, but I'm not
sure this helps his arguments.

The real issue here is that some problems, and especially socially impactful
ones, carry high personal risks if an inference is wrong (problem 1). Therefore,
in order to minimise that, large amounts of potentially sensitive data may be
required (problem 2) and in most, if not all, cases there is a strong argument
for transparency, accountability and human-in-the-loop that makes such
propositions much more expensive and challenging from both a technical (problem
3) and governance (problem 4) perspective. 
Bias poses no doubt massive challenges (problem 5), not just in its
identification and removal, but also in its very definition, as what turns
common sense into bias is a function of our set of social and cultural values,
which greatly varies in time and space. Should history be changed because it
doesn't fit our Western 21st century view of the world? I don't have an answer,
but at least I'm asking the question.
Also, there is the topic of regulation which can't be ignored in these cases:
who establishes the boundaries? Who is it meant to enforce them? Do we need a
universal legislation, or would locally applicable frameworks suffice? Who makes
a call on what is an acceptable risk vs potentially beneficial outcomes -
supranational bodies, governments, individuals, etc.. ? (problem 6)

So the better discussion to have is around the risks involved in deploying AI in
socially sensitive fields, what is the minimum level of benefits it must bring
about to make it worth us facing the challenges above, and how do we go about
having that conversation.

Just branding the whole thing as snake oil is sidestepping the issue and largely
unfair, because it's bypassing the actual conversations to be had and going
straight for the outcome that Prof Narayanan - in arguably a biased opinion -
has decided is what's best for us all.

Artificial Intelligence

We need to talk about AI

It may be argued that customers' satisfaction is the only and ultimate KPI that
really matters when judging the success of a product. In which case Tagmatic
would pass with flying colours, given the fantastic feedback we are receiving
from clients. 

Remarkably enough, we found that sheer inferencing performances are often at the
edges of our clients' considerations: this is partly due to the intrinsically
tailored nature of Tagmatic's models, but mostly attributable to a plethora of
invaluable qualities that go beyond the naked F1 scores: being able to quickly
spin up a model on your own content and vocabulary, and then interact with it
via a minimalistic HTTP interface, zero maintenance, on-the-fly learning of new
classes, low latency, security, scalability and high availability out of the
box, are crucial factors to enterprise-scale deliveries. 

Your Intellectual Property lies in your content, your data, your data models,
and how you use the classifications. It is not in commoditised compute
solutions.
Your resources are thus more effectively spent developing and exploiting your
market differentiating IP, rather than spending way too much time and money on
trying to replicate the engineering complexities of off-the-shelf MLaaS
solutions which may not be your technical team's core competence.

This is where Tagmatic comes in. The hard work and complex engineering has been
done, just bring your content!


The datasets
RCV1 is arguably the most popular dataset for multi-label classification,
comprising of 800,000 newswires by Reuters tagged with 103 classes. Oddly enough
for such a totemic dataset, the breadth of the vocabulary is anachronistic when
compared to some of our clients, modern global publishers who routinely maintain
taxonomies with thousands of topics.
The dataset is available here
[http://www.ai.mit.edu/projects/jmlr/papers/volume5/lewis04a/lyrl2004_rcv1v2_README.htm] 
although unfortunately not in its original XML format with full HTML markup,
which we would normally leverage as part of the classification task. Also the
corpus comes already tokenised and train/test datasets pre-canned. Besides
taking away part of the fun, this has also negative implications on the
vectorisation logic Tagmatic implements, making it largely ineffective
unfortunately. For the purpose of this benchmark, we have used the full train
set (23,000) and a reduced test set (100,000).

AAPD (Arvix Academic Paper Dataset) is a collection of research papers covering
a number of subjects organised in several classification systems. The dataset 
here [https://www.kaggle.com/neelshah18/arxivdataset] is annotated with a number
of classification schemes, of varying level of quality. We have carved out a
subset corpus in the Computer Science domain, comprising of 32,000 papers and 39
subjects in total.


The benchmark
We have considered a number of models amongst the most widely employed for
multi-label classification tasks. The reference evaluation metric is
weighted-F1.

RCV1 AAPD One-vs-Rest - LinearSVC 0.78 0.71 One-vs-Rest - SGD 0.74 0.62 
One-vs-Rest - XGBoost 0.75 0.64 Classifier Chain - LinearSVC 0.79 0.71 Label
Powerset - MultinomialNB 0.55 0.55 Label Powerset - RandomForest 0.60 0.55 
Tagmatic 0.80 0.74 
Tagmatic offers out-of-the-box improvement over a number of basic
implementations. Granted, a team of Data Scientists might possibly come up with
better optimised flavours of such models, pushing performances up. But at what
cost? And by how much? And what then? There is still the much bigger task of
bringing that model to fruition in a real world environment, requiring
non-trivial development effort. Is that really worth it? 
Unless text classification is either your USP or a strong market differentiator,
then the answer ought to be a resounding no. 

Tagmatic is production-ready, enterprise-grade, and available to you right now
[https://datalanguage.com/tagmatic].


Lastly...
We have also benchmarked Tagmatic against Hansard data: a corpus of ~200,000
parliamentary interventions for the 2005/6 period , classified with ~7200
categories, scoring a whopping 0.83. More on this on the next blog post. 
Make sure you sign up for our newsletter [ https://datalanguage.com/contact-us] 
and follow @Tagmatic [https://twitter.com/tagmatic] on Twitter for all the
latest.

Tagmatic

Benchmarking Tagmatic

One the most important tasks for software architects in the early stages of any
technology delivery program is to reach “architecture conceivability”, or
consensus agreement with senior developers that the technical proposition could
actually conceivably be workable. Starting from a position of defensive
scepticism, it typically entails an outside-in narrative exploration of the
potential end-to-end program execution flows at a level of detail between the
Component and Code levels in the C4 Architecture Model [https://c4model.com/] 
for key use cases.

The granularity of the discussion will switch depending on the level of
technical risk entailed by the component. If it is providing a standard
capability that is self-evident to to those present how to code, then there is
no point doing the more detailed code discussion. If it is encoding anything
with a non-trivial level of technical risk, then dropping down to the next level
will be required until everyone is roughly on the same page about how it needs
to work. In this way it acts a bit like a code preview (as opposed to a code
review), functioning in the same way that a delivery prespective
[https://www.malotaux.eu/nrmc.php?id=prespectives] is more efficient than a
retrospective in terms of mitigating risk.

Once and only once this has been completed, you will have established baseline
viability for the program. The narrative described will be unquestionably wrong,
inefficient, full of flaws – that’s not the point. The point is that it will
represent a first, stuttering, throwaway sketch out which a simple, resilient
system design can eventually be evolved. Gaping holes and pointless layers of
cruft will already be visible. Beyond that, most of the selective pressures
driving that evolution will be created later by writing tests and the feedback
from coding itself. 

That’s all fine. However, until you’ve established architecture conceivability,
in reality you’ve got nothing. It is hard to over-emphasise the importance of
this point. Context and Container-level views are meaningless slideware (or at
very best, vague indicators of business process alignment) until they are backed
up by Component and Code-level views. That is solely how they gain their value,
by proxy as maps to decomposing and navigating the the Code-level views of the
platform. Without the code, they are essentially maps to nothing.

Technical Architecture

Architecture Conceivability

In this post we are going to look at some of the ways in which AI and data
science is impacting the practical delivery of technology solutions today. To be
clear from the outset, this will not include a discussion of future-looking
topics that have been getting coverage recently such as fully automated code
generation or algorithmic product development, etc. Rather, we will be looking
the impacts which are already happening right now, the consequences of which are
often just as significant albeit in different ways.

Strategic Alignment Model
To get a clearer idea of what is going on, it will be helpful to start by
reviewing the dramatic shift that has occurred recently within the discipline of
data engineering. In order to do that, we are going to use a version of Niel
Nickolaisen’s Strategic Alignment Model
[http://www.accelinnova.com/strategy.html], which is one of our most frequently
utilised tools in the inception phases of data programs.

It splits an organisation’s capabilities in two core dimensions:

1.     Whether or not they are market differentiating

2.     Whether or not they are mission critical core competencies

This then informs technology delivery strategy as follows:

·      Systems which deliver business capabilities which are mission-critical
market-differentiators you are going to need to build - as by definition there
will be no commoditised solution available yet otherwise it wouldn’t be a
differentiator - and you are going to need to build them very well whilst
effectively managing risk and uncertainty, as this is where your future revenue
streams are coming from (as a side note, this quadrant represents almost all of
our consulting work).

·      Systems which deliver business capabilities which are mission-critical
but not market-differentiators should be bought in. That’s because they are
generally commoditised solutions to common problems (ERP, payroll, invoicing,
etc), hence re-inventing those solutions is a pointless waste of money. In some
cases, organisations might have non-standard workflows which don’t match
commodity solutions, but in the long run it is almost always better to change
those workflows rather than build bespoke software.

·      Systems which deliver business capabilities which are not
mission-critical but are market-differentiating are generally good candidates
for partnerships, i.e. with businesses for whom the newly developed capabilities
will become core competencies.

·      Systems which deliver business capabilities which are neither
mission-critical nor market-differentiating should generally be avoided, as
their value is as best highly questionable.

Such analysis frequently provides very useful insight to organisations. However,
a critical point to remember is that markets are dynamic, so such
categorisations are never fixed. Indeed, it is interesting to see how
technologies traverse this diagram over time. The most common transition is from
market-differentiator to non-market differentiator as technologies mature: new
features and capabilities that were once enough to attract new customers and
generate additional revenue become widely adopted, turning into commodity
features that are required by any product attempting to enter the market or
protect its existing revenue streams. There are also examples, such as online
video within media organisations, which start out as non-mission critical
differentiators before becoming mission-critical differentiators and then
non-differentiating core competencies (as a side note, the other important shift
as functionality matures from differentiating to non-differentiating is the
associated switch in customer priorities from functional to non-functional
requirements – what matters for non-differentiating systems is
standards-compliance, portability and the avoidance of vendor lock-in, etc).

Data Engineering As Market Differentiator
So what does this have to do with data engineering? Well, the really interesting
thing about the recent history of data engineering is that is represents the
much less common example of a technology moving in the opposite direction –
upwards, from non-differentiating to differentiating:

Data engineering used to be the domain of the back office: of ETLs, PL/SQL
programming and cron jobs. It was one of the areas of software development most
insulated from the market pressures which shaped agile software delivery
techniques and lean product development. It was also one of the least impacted
as a result: rare indeed was it to find an ETL developer writing their code
test-first.

However, the revolutions in firstly data analytics and statistical software
development, and then in machine learning techniques, have radically transformed
the discipline and moved it front and centre in the new era of data-driven
product innovation. This has created a sudden critical demand firstly for access
to these new technologies, a need which has driven the recent adoption and
growth of data science as a commercial capability; and secondly for the software
development skills that allow innovative product ideas to be tested and iterated
rapidly under conditions of uncertainty without incurring crippling amounts of
technical debt.

One of the most prevalent organisational responses to these demands is for
data-driven product delivery programs to employ three discrete technology
functions:

1.     Data science teams: typically, recent PhDs with good theoretical
knowledge but little experience of commercial software development. They often
work in what is effectively an R&D function developing prototype implementations
in their preferred language.

2.     Software delivery teams: where these function best, they are most
typically staffed by experienced/senior developers well versed in extreme
programming and continuous delivery techniques and ideally who have worked
together previously (the importance of this last point is so often overlooked –
no matter how good the individual members are, if the team itself is new then
you are going to burn significant time and cash whilst relationships, trust and
delivery flow are negotiated and established
[https://en.wikipedia.org/wiki/Tuckman%27s_stages_of_group_development]).

3.     Machine learning engineering teams: who are responsible for taking the
outputs of data science and adapting them into something which can be integrated
by software delivery. I use the term “adapting” rather than “refactoring” here
intentionally, as due to the “specialised” nature of the data science algorithms
the aim is normally to add runtime tooling into a software execution wrapper
that allows the prototype implementation to be invoked concurrently or to handle
larger datasets. Most typically this will be suites of map-reduce and derivative
tooling (not least because these teams are also the most natural home for data
engineers who invested in Hadoop stack upskilling during the first data
analytics revolution).

Implications for Software Architecture
This differentiation is having profound, and most highly undesirable, results.
My colleague Fabio Colasanti has previously written persuasively
[https://datalanguage.com/features/data-science-is-about-building-software] 
about the consequences at the code level. In the rest of this article, we will
explore the implications for software architecture, and the resulting impacts on
the business value actually being realised by these programs of work.

Arbitrary Service Boundaries
Conway’s Law [http://wiki.c2.com/?ConwaysLaw] states that the architecture of
systems tends to reflect the architecture of the organisations that build them.
The most reliable way to mitigate this tendency is to create a single
cross-functional delivery team that includes domain experts and works as close
to customers as possible. That’s because the domain model and business
capabilities encoded into software create the channels along which information
flows and business value is then generated and delivered.

In the absence of such an integrated, cross-functional approach, the likely
outcome will instead be an architecture where service boundaries reflect the
arbitrary divisions of labour in the construction of the codebase. Is that
really so bad? Well, yes actually - it is. Once the code sequences required to
implement and integrate the business capabilities are defined, it becomes
immediately apparent that the natural execution flow is being critically
disrupted by meaningless service interfaces that require mapping layers,
additional infrastructure complexity, kill performance, undermine cacheability,
expose security vulnerabilities, and introduce a host of other avoidable
problems.

In doing so, they also undermine the chances of a successful delivery.
Successful technology delivery is ultimately a question of risk management: of
identifying risks and then managing or mitigating them so that their cumulative
impact is as low as possible. That means ruthlessly removing all avoidable
sources of risk and noise - so that the inherent risks of the market (changing
customer demand, competitor behaviour, etc) can then be observed, responded to,
and their associated opportunities exploited.

Self-generating Complexity
In response to the implementation barriers created by arbitrary service
boundaries, delivery teams then typically attempt to work around them via the
lowest friction route available. Rather than conducting a politically
unpalatable root cause analysis and proposing a simplified approach that
conflicts with the structure of the overall delivery program, it is easier to
try using additional software and infrastructure complexity to address the
problem. This results in an escalating feedback loop of bad design: not only is
the architecture unnecessarily complex already, but additional complexity is
then introduced to manage that complexity.

Unfortunately, the only way to break out of this loop is for some brave soul to
put their hand up and conduct a “survival of the fittest” design cull, where
every component of the architecture is ripped out and tested for critical system
viability until what remains is the minimal skeleton actually required to
deliver the baseline functionality.

Machine Learning as Universal Panacea
A more subtle variant of self-generating complexity relates to the application
of machine learning itself. Given the relative immaturity and vast potential of
the field of data science, there is now the additional lurking temptation to try
using ML to work around any problem rather than analysing and solving its root
cause. However, as Robert Frost said “the best way out is always through” – a
phrase which captures the essence of how to avoid incurring crippling technical
debt. Applied to data-driven product innovation, that means conducting patient
and precise analysis to identify underlying causes of problems and solving them
at source rather than building an ever-expanding shanty town of workarounds that
is liable to catastrophic and terminal failure.

Effective Mitigation Strategies
What techniques can be employed to mitigate against these problems and create a
reliable foundation on which to base your delivery? In our experience the two
most valuable tools are a domain model and sets of BDD scenarios (or terse use
cases) which describe the state and behaviour on which the anticipated business
value depends:

Domain Model: this is a collaboratively defined conceptual model of the problem
space shared between the subject matter experts and technical staff. Its key
importance is that it explicitly names the entities of the business domain and
their relationships to each other. In doing so, it creates a worldview and
language that is shared by all stakeholders/members of the delivery team (a
purposefully limited and relativistic scope, which is enough to ensure use cases
can be fulfilled whilst avoiding the many pitfalls and dead-ends of enterprise
data modelling). It is not a knowledge graph, ontology or database schema.
Rather it is something which SMEs might express as a frequently iterated
whiteboard diagram of labelled entities and connectors, and developers might
describe using an ERD or high-level class diagram. This foundation is what
allows code to be written in a way that precisely and unambiguously matches the
collective intent of the team. Without it, you are building on quicksand.

BDD Scenarios: once a shared language has been established which describe the
entities in the domain and how they relate to each other, the state transitions
involved in the information flows and generated business value can then be
captured (in practice this tends to be highly iterative, with new scenarios
being sought that break the model and thereby drive its improvement). The
emphasis here is not on automated testing capability – although that can be a
very helpful by-product - but the communication of state transitions and
workflows [https://lizkeogh.com/2013/10/24/bdd-before-the-tools/]. 

Once the workflows and information flows have been defined via which business
value will be delivered, questions concerning the optimal deployment of human
and machine agents can then be correctly and meaningfully addressed. At this
point, it is then really important to return to the Strategic Alignment Model
for one more pass. Newly equipped with a powerful model that clearly describes
the state/data and behaviour of your domain, you can review each entity and
function in turn to create an unambiguous and precise definition of what your
market differentiators really are. Is it really the bespoke algorithm that your
data scientists are working on, or is it in fact the data that the algorithm is
outputting? If the latter case, is it really justifiable to keep investing in
developing something in-house which ought to be (and will almost certainly end
up eventually) a commodity purchase? These considerations will then be primary
factors informing product roadmap planning – a subject beyond the scope of this
post.

The result should be a system architecture that is the simplest thing possible
[https://www.jamesshore.com/Agile-Book/simple_design.html] to deliver the
identified business value, and as a result is the cheapest to build, contains
the minimum number of things that can go wrong, and is fastest to adapt to
future market change.

Current impacts of AI on Technology Delivery

Silver Oliver

An introductory guide to tagging: part 4

In the previous post I talked about the challenges of tagging
[https://datalanguage.com/features/why-tagging-tends-to-be-bad-on-websites].
Here I will start to look at some approaches that can help.

One technique that I will always come back to is domain-driven design (DDD). The
principles behind domain-driven design go back to Eric Evans and his book 
Domain-Driven Design: Tackling Complexity in the Heart of Software
[https://www.oreilly.com/library/view/domain-driven-design-tackling/0321125215/]
.  The starting point for domain-driven design has similarities to domain
analysis in Library Science. 

In domain analysis Albrechtsen and Hjørland make the case for domain knowledge
being given greater importance in the role it plays in Information Science.

> Many approaches to information science and Knowledge Organisation may be
understood as attempts to pass over subject knowledge (or at least not make
subject knowledge explicit in their methodologies). Domain analysis, on the
other hand, makes subject knowledge an explicit and important part of the
methodologies of information science and knowledge organization.
Towards a new horizon in Information Science
[https://www.researchgate.net/publication/220434297_Toward_a_New_Horizon_in_Information_Science_Domain-Analysis]

DDD, though aimed at software developers, comes to a similar conclusion.  This
is that many software projects fail due to passing over subject knowledge and
not making understanding the domain a central part of the process. The problem
being that everyone on a team goes away and builds according to their private
view of the domain. Then when it comes to putting the design pieces together
they don't fit.

DDD ensures a team spend time together at the outset of a project to align their
understanding.  This is before they design or build anything.  The success of
this approach is that it supplements existing expertise with a strong and shared
knowledge of the domain.

The aligning of language during these sessions is crucial.  Participants talk to
each other, draw for each other and do everything they can to understand where
differences of understanding or terminology are lurking. 

Diagrams and word documents have the habit of hiding our differences as opposed
to surfacing them. True shared understanding is something that can only really
be achieved by talking about the domain from as many angles as possible. 

It is as simple as that. 

When starting on a project to design and build anything, I would suggest the
following principles:

 * Recognise that caring about and understanding the subject domain is
   important.
 * Domain experts are a critical part of the team.
 * Everyone will have different views, preconceptions, simplifications they
   bring to a domain and this manifests itself in the language they use.
 * Learning where these differences are will be expensive if learnt through
   writing code,  building a taxonomy or designing a user interface. Better to
   learn this cheaply in front of a whiteboard.
 * The best way to surface these differences and align language is in talking, a
   lot, together as a team.
 * Simple domain model diagrams are a great way to share and challenge
   understanding of the domain.
   

The main outcome of all this is the shared language during conversations and in
design decisions.  There will likely also be some domain model diagrams.  An
example of this is the classic Polar Bear presentation
[https://www.slideshare.net/reduxd/beyond-the-polar-bear/22-stadiumbr_competitionbr_teambr_matchbr_roundbr] 
by Mike Atherton of a sports website. 


Credit Mike Atherton [https://twitter.com/mikeatherton] and Paul Rissen
[https://twitter.com/r4isstatic] (model design)This represents the project
team's shared understanding of the domain and consensus regarding how to talk
about it. In this model are all the types of tags we might need. As we
understand more about the website's goals and the plan for which pages are to be
built first we can ensure tagging will support them.

For example we might decide we want to capture tags for teams and matches. Why
not stadiums? If we tag with a match we already have the stadium from the model.
 So we can reduce the burden on taggers to that which is absolutely necessary.
We can also see we get competition round from the model as well. Each tag is
part of a larger model and we adjust our tagging strategy based on our
understanding of the model.

Why not tag with player?  Analysis may show that there is little value in
including players. For example we have seen little use of player pages in other
parts of the site. Then there is no justification in tagging them. Some might
argue that we need to preempt any possible use of the system in the future, so
we should tag players even though this doesn't add value now.  In my experience
this is where we get into trouble.  Keeping a tight focus on the value and the
tagging needed to create that value is crucial. The domain model will help us
map out possibilities but an understanding of our goals helps up pick the right
path through that map.

To continue this example, if we identify that tagging with teams and matches
will  support the objectives of the site then we have an idea of what good looks
like. This informs how we develop the tools, vocabularies and guidelines to
support this goal. An understanding of the domain, the websites goals and its
evolving scope have defined how we design our tagging.

That is a simple example but there are a whole range of more complex case
studies to support this from Wellcome
[https://stacks.wellcomecollection.org/building-a-free-and-unrestricted-digital-museum-and-library-6cdf28784068] 
, BBC [https://www.bbc.co.uk/blogs/radiolabs/2009/01/how_we_make_websites.shtml] 
and Parliament
[https://pds.blog.parliament.uk/2016/09/09/modelling-parliaments/].

I often get asked if this goes against user-centred design philosophies? Yes and
no. 

When working at its best DDD is complementary and works hand in hand with user
centred approaches.  At its core is the idea that the domain as a key organising
principle.  This can still sit within a user-first approach.  However, on the
whole I feel a team aligning its understanding and language within a domain
before engaging with users can only be a good thing.

In summary, one part of the solution to better tagging is to design with a
collective understanding of your systems domain. 

Part 1: Why the web turned its back on the librarians. And why we need them
back.
[https://datalanguage.com/features/why-the-web-turned-its-back-on-the-librarians]

Part 2: The importance of tagging when publishing on the web
[https://datalanguage.com/features/the-importance-of-tagging-when-publishing-to-the-web]

Part 3: Why good subject tagging is hard
[https://datalanguage.com/features/why-tagging-tends-to-be-bad-on-websites]

Domain-driven Design

How to use domain modeling to improve subject tagging

An introductory guide to tagging: Part 3

It has been my experience that many organisations struggle with high quality
tagging. In this post I am going to reflect on some of the causes of this
problem. 

As discussed in earlier posts, the value of tagging is often not properly
understood
[https://datalanguage.com/features/the-importance-of-tagging-when-publishing-to-the-web]
.  Most often it is considered to be there to support keyword search alone.
 Even the taggers themselves are not aware of how the tags are going to be used,
or why they need to be of good quality to add value.

The next issue is around the guidelines given to taggers.  Typically they are
extremely poor, and go along the lines of “tag what the content is about”.  This
question of “what is it about?” is harder than it seems at first glance and not
very helpful. It usually results in a different answer from each tagger.  To get
good quality tagging you need to ask good quality questions.

As the tagging quality goes down, the ability to drive useful features from the
tags also diminishes. This in turn does nothing to reassure people of the
importance of good tags and you end up in a vicious cycle of ever decreasing
quality and value. 

What good guidance on tagging looks like should be the point where Library
Science can come to the rescue.  But the challenge of how to identify the
subject of a work is a surprisingly under represented topic in Library Science
literature. 

Back in the the 1970s, Hutchins observed:

> THE LITERATURE OF INDEXING AND CLASSIFICATION contains remarkably little
discussion of the processes of indexing and classifying...we find very little
about how indexers and classifiers decide what the subject of a document is, how
they decide what it is 'about'.
The concept of 'aboutness' in subject indexing:
[http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.584.8116&rep=rep1&type=pdf] 
 W. J. Hutchins

Little has changed.  Taggers are often told to imagine "how might the user
search for this piece of content" or "take the author's view and characterise
what they are trying to say".  Neither is very useful as a framing question. 

Hutchins, in his paper, tackles this problem and introduces a distinction
between summarisation and aboutness. Summarisation is the representation of the
total subject content of a document and the statements made therein. While
aboutness is about capturing the thing that has been written about.

A very simple example might be a sport story about a player and whether they
will move clubs, their struggles with injury and their performance this year.
Many arguments are developed, making up the narrative of the story.
Summarisation would capture not only the player as a tag but also attempt to
capture something about the claims made about the subject of fitness,
performance and likely future at a club.  On the other hand, aboutness only
attempts to capture the thing being talked about and not what claims are made
about them. In this case just the player.

This distinction is rarely made in tagging guidelines and the tagger is left to
balance the level of summarisation and aboutness on gut feel.

Hutchins goes on to characterise this distinction in another interesting way.
The aboutness represents those things that are expected to already be part of
the readers knowledge. Every piece of content assumes a level of knowledge of
the reader with which the narrative then builds. Summarisation on top of this
attempts to capture what is new and builds on that existing knowledge in the
narrative of the content.  The readers are familiar with the player, but not
about the subtle changes in his fitness and future prospects.

With this in mind, the questions we might ask a tagger to answer are: What is
the thing being talked about? What claims or facts are made about a thing in the
content? What might the reader be presumed to know before reading this?

Whether we choose to do summarisation or capture aboutness is a decision related
to the desired outcome of the system. But we should be clear about which one it
is.

A more recent analysis by Birger Hjorland (The concept of subject in
Information
Science [https://www.emeraldinsight.com/doi/abs/10.1108/eb026895]) revisits the
problem.  He cites many examples of Guidelines expecting taggers to do little
more than perform "mind reading" on prospective future consumers of the content.

Instead, he argues, we must think about tags as a tool in enabling the full
potential of the document within the context of the overarching system.  The
desired outcomes of this system should then dictate the tagging strategies and
guidelines. 

It is this domain and system that must be crystal clear in taggers' minds as
opposed to the idea of aligning subject to a theoretical searcher.  They need to
also understand the value that they are creating and how.

Sadly, this is the opposite of what we see in many tagging tools and guidelines
today.

I would argue that tagging should be a contract between taggers and the system
 much in the way we see in the design of data services
[https://medium.com/@diogo.lucas/https-medium-com-diogo-lucas-contract-breaks-are-kinda-of-deal-breakers-447762de20c4]
.  Tagging guidance needs to be managed with a vision of the desired outcomes
combined with asking clear and simple questions of the taggers. This is as true
for humans as machines.

In summary, tagging outside of libraries and archives is often of poor quality.
Value is rarely understood, content creators are often asked to tag in ways that
are a chore and cognitively awkward. Worst of all, the question of what good
tagging looks like has rarely been asked.

In the next post I will look at some potential solutions to the challenge of
better tagging
[https://datalanguage.com/features/how-to-use-domain-modeling-to-improve-subject-tagging]
.

Part 1: Why the web turned its back on the librarians. And why we need them
back.
[https://datalanguage.com/features/why-the-web-turned-its-back-on-the-librarians]

Part 2: The importance of tagging when publishing on the web
[https://datalanguage.com/features/the-importance-of-tagging-when-publishing-to-the-web]

Part 4: How to use domain modeling to improve subject tagging
[https://datalanguage.com/features/how-to-use-domain-modeling-to-improve-subject-tagging]

Blog