Getting started

in project data
 Contents

Acknowledgements.......................................................3
Introduction.........................................................................4
Imagine..................................................................................5
Who’s the guide for?......................................................6
What is project data analytics?.............................7
What does it cover?........................................................8
What are the opportunities?....................................9
Making a tactical start................................................10
Taking a strategic approach..................................12
What tools help...............................................................15
Fequently asked questions......................................18

2 Getting started in project data

 Contents

Acknowledgements

This collaborative report is part of our research

collaborations. Research collaborations seek to bring
together like-minded individuals and organisations to
develop and generate innovative research that supports
the needs of practice whilst helping to promote a
profession built around learning and collaboration. For
more information, please visit apm.org.uk/research

Contributors

Alex Robertson, Petrofac

Andy Murray, Major Projects Association

Gareth Parkes, Sir Robert McAlpine

Martin Paver, Projecting Success

Reviewers
Daniel Nicholls, APM

Eleni Papadonikolaki, UCL

Grant Mills, UCL

James Dowle, PA Consulting

John McGlynn, Atkins

3 Getting started in project data

 Contents

Introduction

We’re pleased to publish this report on behalf of the Data More information on the Data Advisory Group can be
Advisory Group and PDA Taskforce. This guide is part of a found at:
wider leading collaboration between the Data Advisory apm.org.uk/resources/what-is-project-management/
Group and the PDA Taskforce which aims to explain how what-is-project-data-analytics/
data analytics can help projects and what you can do to
get started. Other collaborative activity includes events,
blogs, and interactive media, as well as a Data Maturity The PDA Taskforce comprises a team of senior level
Model from our partners. professionals who have a mission to use data to transform
project performance with a focus on delivery and action,
working in the interests of the collective rather than
What is the Data Advisory Group and PDA any organisation or sector. The team is made up of a
Taskforce? broad cross section of expertise covering: public sector,
private sector, academia, and a membership body; large
We have convened a Data Advisory Group to bring corporates through to SMEs; clients through to vendors;
together professional bodies including APM and the Major project delivery and data professions covering several
Projects Association (MPA); government including the industry sectors. The team has been structured to include
Infrastructure and Projects Authority (IPA); academia an ideal mix of expertise to drive forward a bold and
including the Universities of Oxford, Warwick, and UCL; aspirational agenda. The team is not exclusive as we
funding providers such as UKRI; industry representatives encourage the broadest possible engagement via our
including Network Rail and PA Consulting and leading workstreams and associated events.
organisations in the field of project data and analytics
such as Projecting Success and Sir Robert McAlpine. The More information on the PDA Taskforce can be found at:
group aims to improve ‘data literacy’ enabling individuals pdataskforce.com
and organisations to understand how they might be able
to make better use of data, supporting and signposting
developments and like-minded initiatives such as the PDA
Taskforce and Project X and sharing learning across the
profession.

4 Getting started in project data

 Contents

Imagine

Imagine a world where data from previous projects data rules that trigger automated actions to your team, This guide aims to explain how data analytics can help
predicts where future projects are likely to diverge from effectively ensuring a standardised minimum expected projects and what you can do to get started.
the plan. Where we understand the return on investment level of performance with little ongoing effort – both for the
and long-term benefits realisation or know where best customer and the contractors. Performance will increase,
to invest contingency. These insights are encapsulated mistakes will decrease, decisions will be quicker.
in the experience we accrue through projects. Often, we
distil this down into selective abstracts and call it lessons The project management community want predictable
learned or codify it in processes, methods, and handbooks project performance and better outcomes. We should
but we all know that there are complex relationships and always know, before sanction and during delivery, the
cause-effects that exist that are difficult to summarise into likelihood of our project being a success, however we don’t
a paragraph, page or book. We can now deploy the latest currently, as proven by academic literature.
methods in data science to find patterns in data that help
us to gain insights that would otherwise have been difficult
to uncover. Data analytics can give project professionals
new superpowers. This may sound far-fetched, but such
uses of data analytics are already being used on projects
and are well established in other domains.

Imagine as a step along this journey you can access all

your project information within three clicks of a mouse,
without logging in to multiple systems or searching
through your emails for an attachment. Imagine fully
automated daily reports and the extra time your site
supervision will have to manage safety and increase
site performance. Imagine, also, that your project
management office (PMO) have set up role-based

5 Getting started in project data

 Contents

Who’s the guide for?

Our guide is aimed at project professionals or senior leaders

of project-based organisations who wish to improve project
delivery on a specific project or across their organisation, e.g.
CEOs, CTOs, PMO lead, project/programme managers, and
also business intelligence directors, directors of education and
heads of talent etc. This guide will also be of interest to the new
roles that are emerging such as head of digital, data leader,
head of data and more.

This guide assumes you’re experienced in project

management and implementing business change, therefore
we don’t cover general change management approaches
for adopting new practices or innovation. Instead, we
focus on what is meant by project data analytics, what the
opportunities are, what’s happening now, and the technology
and practices being used. We also aim to debunk some myths.

This guide will help you understand opportunities for

improvement made possible through project data analytics
and how to get started. This will include tactical activities to
generate momentum and strategic activities to maximise
potential. It will also help you avoid pitfalls, reinventing wheels
or racking up unnecessary costs from advisors or solution
providers.

Project data analytics will impact every project delivery role,

not just project management. This guide has application
beyond the traditional boundaries of the APM and is intended
to encourage collaboration across professions.

6 Getting started in project data

 Contents

What is project data analytics?

Project data analytics is the use of past and current project data to enable effective decisions on project initiation,
delivery and efficient automation of project tasks. This can be done in three ways:

• Firstly, by using algorithms to respond to data inputs to perform routine tasks faster with fewer errors than if
processed manually;
• secondly by collecting, cleaning, analysing, structuring and presenting data in the most effective format to make
better decisions;
• and thirdly, by using it to predict future performance to make timely decisions and actions.

Project data Analytics

Projects generate lots of data from the day they’re Analytics is the process of using data to enable
first conceived to many years after they’ve finished, effective decisions or efficiently automate tasks. In
and impacts/benefits are assessed. Data may range this respect, the project world has been applying
from content in project documentation, project rudimentary project data analytics for a long time, be
correspondence, project operational systems, or that reviewing project dashboards, holding lessons
project logs and records to name a few – for example learned workshops or running Monte Carlo analysis of
data in a project schedule, data in a supplier contract, risk registers. However, until recently such analysis has
and purchasing records or minutes of project been limited by the high cost and low performance
meetings. Current technological advancements of data collection, data cleaning and processing
and data infrastructures have contributed to an capabilities. Today, the rapid cost-performance
increase in historical data that could be used for improvement in data processing combined with
performance monitoring. This data may be used to advances in data science means the handbrake of
inform decisions relating to the project that generated such limitations is off. We can now move from making
it or it may be used by the host organisation, industry decisions based on manipulated and out-of-date
associations or academics to inform decisions on data which is limited to a handful of datapoints to
other similar projects or on methodologies and project making decisions on near real time source data
management practices generally. across large volumes of datapoints. We are moving
from guesswork derived from anecdotes to genuinely
data driven decisions.

7 Getting started in project data

 Contents

What does it cover?

Project data analytics covers a wide range of capabilities. Application Programming Interface (API) Artificial Intelligence
Whilst artificial intelligence (AI) grabs the headlines,
that’s just one form of data analytics. On a scale of least An API is a connection between computers or between In a project context, AI systems could advise on or even
intelligent/autonomous to most intelligent/autonomous software applications. It’s a type of software interface, make project decisions, which would certainly enhance or
we have robotic process automation, machine learning offering a service to other pieces of software. APIs are change the roles and status of many project professionals.
and artificial intelligence. They are very different, as needed for allowing different applications, especially Narrow AI is created to solve a specific problem, such as
described below. legacy systems, to talk to one another. APIs and extracting recommender systems or chatbots. General AI does not
data from disparate sources are increasingly becoming rely on human programme to accomplish tasks; it will
part of the job of data analysts. need to understand context and environment and adapt
Robotic Process Automation (RPA) accordingly. Although General AI may still be decades
away, Narrow AI will increasingly dominate project delivery
Uses technology to automate business processes by Machine learning based use cases. Solutions are beginning to proliferate;
applying business logic to process inputs (e.g. receiving low code capabilities make them increasingly accessible
an invoice or a progress update). Using RPA tools, an RPA is transactional by its very nature and therefore to all of us. Our challenge will be how we integrate these
organisation can configure a software robot (or ‘Bot’) to limited to the business logic that is configured for the capabilities and standardise, whilst retaining a degree of
control computers to process a transaction, manipulate bot – so unintelligent processing that won’t handle control over the data.
data, trigger responses and communicate between unexpected input and won’t adapt as context changes.
digital systems. RPA scenarios range from something as Whereas, Machine Learning, is intelligent processing in that
simple as generating an automatic response to an email it involves computer algorithms that ‘learn from doing’. Other digital innovations
to deploying thousands of Bots, each programmed to For example, when reading a form, such as an invoice,
automate jobs in a system such as a project management RPA Bots will be told where to look for items such as the There are lots of other digital innovations which will
tool or finance system. With RPA, businesses can automate date, the supplier’s name, invoice number, and amount. If improve project delivery e.g. blockchain technology
mundane rules-based business processes, enabling these are in various places or in different formats it may applications such as smart contracts, use of augmented
individuals and teams to devote more time to performing not read them properly. With machine learning, the can be reality in design development and stakeholder
higher-value work. In a project context RPA can be used for trained to understand the characteristics of these items engagement, use of drones for quality control or earned
regular routine project tasks where predetermined actions and therefore detect them regardless of where they are value on construction sites, Building Information Modelling
can be defined on the vast majority of expected inputs or if they’re in different formats. This process gets more (BIM), Digital Twins, Internet of Things (IoT), etc. This guide
(e.g. producing project status reports, updating project accurate the more it’s used or trained. In project terms, is focused on data analytics therefore we will not cover
finance records). It’s like workflow but without manual machine learning, can be used to spot patterns between these other digital innovations.
entry or manipulation. some characteristic of projects and some aspects of
project performance to provide insights to aid decision-
making.

8 Getting started in project data

 Contents

What are the opportunities?

One of our recent reports estimated that project Project data permeates all aspects of project “Better use of data in infrastructure project delivery
management creates between £150-170bn of economic management and project delivery, so the opportunities for could save £23bn a year. At root that’s about tackling
value annually in the UK 2. Yet, successful project delivery in project data analytics cover the full spectrum of project poor decision making. We now have more data than
terms of effectiveness and efficiency remains stubbornly activities. If you’ve ever uttered the phrase “if only we had ever on our projects to help us make better decisions
low. Analysis by the University of Oxford’s Bent Flyvbjerg known” or “I wish we had done X instead” then it’s likely – but project data analytics isn’t widely understood or
and Alexander Budzier on ~12,000 global major projects that data analytics may have helped. adopted. We need to understand why and what can be
concluded that the probability of a project being delivered done about it.”
on budget, on time, and delivering the full benefits defined Data analytics can help with project level decisions
in the business case is only 0.5% 3. That is only one in 200 such as understanding changes in stakeholder Naomi Brookes
projects delivering on all their promises, so the scope to sentiment, assessing productivity of individuals and University of Warick
improve the initiation and delivery of projects is huge, as is teams or understanding supply chain risks. It can also
the scale of the economic impact. help at portfolio level such as understanding project
predisposition to succeed/fail and therefore which 2
Association for Project Management (2019). The Golden Thread – A Study of
So, how can project data analytics help improve project projects need the most leadership support or how many the Contribution of Project Management to the UK’ Economy and Society.
3
oxfordglobalprojects.com/portfolio and parliament.uk/business/committees/
performance? If a project was a manufacturing process projects are affected by socio-economic or environmental committees-a-z/commons-select/public-administration-and-constitutional-
it would have hundreds, if not thousands, of sensors trends or risks. affairs-committee/inquiries/parliament-2017/inquiry1/publications/
generating performance data to be analysed in near real- 4
rics.org/globalassets/rics-website/media/news/news--opinion/rics-future-
report-2.pdf
time with the ability to spot trends, model options, optimise The cost-performance improvement of analytical tools
controls and maximise yield. However, the typical project and advances in data science means that solutions
dashboard comprises around 10-20 data items reported a are readily available and viable for nearly all projects –
week or two after month end. Indeed, a recent RICS report 4 some tools may even be already part of your enterprise
claimed that 95% of project data is not even used. Projects license. So, don’t feel that data analytics is out of reach
therefore need to make the shift from being managed as or would take too much effort to make it worthwhile. We
temporary initiatives operating outside of business-as- have included some examples to show what’s currently
usual structures (resulting in slow and manual collection possible.
and manipulation of data) to being treated more akin to
an operational process that routinely collects and uses
performance data. The data is there, it’s just that we need
to establish the mechanisms and habits to collect and
analyse more of it, more often and from the start.

9 Getting started in project data

 Contents

Making a tatical start

Firstly, don’t put off getting started. Many people will Identify opportunities to deploy solutions on a pilot Have a look at those as there are likely to be some
have already done some form of data analytics, but basis. Identify and develop advocates, working with candidate issues/opportunities where data analytics can
without necessarily knowing it. This could be a macro in senior project leaders to facilitate data access and help including:
a spreadsheet, a rule to process emails in your inbox or a experimentation. Leverage existing community driven
morning routine with your smart speaker at home. Project solutions such as from Project:Hack, tailoring to your • Reporting - transitioning from PowerPoint reporting to
data analytics can build on such macros, rules or routines. business rather than starting from scratch. real time dashboards, with integrated supply chain data.
• Automation - from simple progress chasing workflows
A challenge for some organisations is that they may Initial solutions at project level could include introducing through to automatically updating disparate corporate
not know what data they own, what they have rights to, some simple automation for aspects of the project systems.
where it is and whether it’s good enough to enable a with high levels of repetitive transactions, e.g. logging, • Risk analytics - how often is the risk register updated?
data-driven approach to project delivery. If that’s your progress chasing, categorising and allocating design Are the updates substantial? How well are the risks
organisation, don’t worry - all organisations have access queries or autogenerating project updates for specific defined? How effective is the management action? Are
to sufficient data to make a start. Project data analytics stakeholder groups. For automation, look for resource the risks what you would expect to see for this type of
is different from big data, so you don’t necessarily need intensive activities which are mundane, are error prone or project?
lots of data and a top-down big budget solution to get take longer to complete than the timeframe that makes • Schedule variance analytics - how good is the project
you started. Project data analytics can be undertaken on them most useful. To help identify initial solutions at the team at estimating? Which work packages are more
very small data volumes such as <100 rows of data in a enterprise level your PMO probably has a backlog of predisposed to variance?
risk register. Such data may only be updated every two improvement ideas that have either not delivered or were
weeks or more, so that data volumes aren’t that high. put on the too-hard pile. Alternatively, speak to internal
Eventually, you could move into big data challenges when audit who will be able to highlight issues they see on a
you interconnect and correlate data, version control it and repeated basis.
feed in data streams for recording project activities. But
for those starting out on project data analytics, you are
unlikely to be grappling with big data type challenges, to
begin with.

There are three tactical approaches. You can look to apply

project data analytics on a single project, or you can apply
it at enterprise level but on single issue (e.g. risk). Or you
could do both! Whichever approach you choose, resist
the temptation to go for predictive analytics initially as
that requires machine learning, underpinned by a more
strategic approach.

10 Getting started in project data

 Contents

Making a tatical start

Secondly, raise data literacy and awareness. What we Some will address this tactically, addressing specific pain “Start small and iterate – use tactical improvements to
have seen is that if you bring in PhD level data scientists points and building momentum. Others will take a more create capacity. Build your understanding and that of
into a project delivery organisation and they spend strategic approach, underpinned by a roadmap and an your team, you’ll steadily see more of what’s possible.
most of their time sorting out data and battling with IT organisational champion with responsibility for making Find and grow pioneers and other leaders, then work
departments, then they’re unlikely to stay around for very things happen. together to develop your ideas to get senior and
long. Although it’s good to take a strategic approach, we strategic support.”
would advocate starting off with awareness raising and For practitioners we recommend attending some meetups
gaining rudimentary skills before moving to a strategic or joining a hackathon, further details of which can be Gareth Parkes
approach to skills development: found at the back of this guide. There are an increasing Sir Robert McAlpine Ltd
number of training providers but also consider using
• Upskill your own team in PowerBI or Tableau and begin Massive Open Online Courses (MOOCs) such as from
to make the most of visualising data using dashboards. edx.org, Future Learn, Coursera or Linkedin Learning for
Ensure that people understand the fundamentals rather upskilling your project teams and PMOs.
than using the capabilities as a glorified charting tool
otherwise it will be difficult to scale up. Thirdly, showcase what’s possible and then make the
• Provide key people who have a passion for the topic with case for a strategic approach. Focusing on tactical
an opportunity to advance themselves via structured improvements on an opportunity/pilot basis, driven by
training. Identify and develop superusers. the advocates within the business who have a passion for
• Explore opportunities to deploy automation. Begin a specific technology enables the organisation to gain a
with no code or low code solutions, but also consider deeper understanding of the capabilities.
Python code-based solutions. The latter avoids licensing
costs but can result in a higher maintenance burden Develop a shared pipeline of ideas, working together to
compared to procuring off-the-shelf automation drive change. Prioritise the pipeline, focusing on solutions
solutions. Try running localised pilots for automation with a tangible business impact that attracts senior
before sharing across projects and departments. management attention and helps to inspire others.
• Mobilise a local community of practice with people who Demonstrate the art of the possible, starting small and
have expressed an interest in upskilling. experimenting with higher-end solutions – making sure
to track the benefits. When you have secured a senior
Some organisations have taken this first step by sponsor develop the case for a more strategic approach
employing third parties who help to illustrate the art of the using your own examples of what’s been achieved so far.
possible and secure buy in at board level. They tend to get
there more quickly, but skills transfer should be considered.

11 Getting started in project data

 Contents

Taking a strategic approach

Taking a strategic approach means taking an enterprise • Governance - will you be allowed to deploy solutions?
approach - a proven way to effectively invest in What constraints will your IT department impose?
capability is by taking a maturity model approach. We • Legacy systems - most organisations won’t be starting
have produced a report that highlights the current use from a blank sheet of paper. Are you locked into
of project data analytics. This section highlights some legacy solutions? Do the systems talk together via an
key considerations of a strategic approach regarding Application Programming Interface (API) or does the
data strategy, data audits, data architecture, skills and data sit in silos and need to be manually collected or
embedding it within your organisational culture. transferred?
• Scope - does the data you capture cover everything
about the project that you need to know, or do you have
Data strategy several suppliers, partners, and clients who each hold
data required to complete the project jigsaw?
Every project delivery organisation should look at
developing a data strategy, irrespective of how immature Underpin the strategy with a maturity model so that you
your understanding of project data analytics is. If you don’t can measure progress. Deploying project data analytics
have a strategy, how do you know where you’re heading? should be treated as a change program, it isn’t simply a
Map out a high-level vision and roadmap for change. This matter of supplying licences for the latest and greatest
is likely to be a multiyear journey that blends short term tools. Our data advisory group are working in conjunction
benefit with longer term improvement and integration with WMG on developing a data analytics maturity model
of data pipelines. The strategy should also capture any which may be able help with this.
constraints. This will include considerations such as:
“You do have to get into the detail. It’s imperative to
• Affordability - does budget exist for additional software, understand the question you want to ask of the data, what
licensing, cloud services, technical support, and training? outcome you want for different users, what processes
• People and culture - how open is your organisation to the data is driving. You also need permission and time to
making data accessible and how will you overcome any explore, test and iterate. That’s how you get confidence.“
barriers? What are the new roles and responsibilities that
are needed? Daniel Byrne
Infrastructure and Projects Authority

12 Getting started in project data

 Contents

Taking a strategic approach

Project management review currently have and the benefit that you want to get from it is it heavily swayed towards large scale deployment? Will the
(or the problem statements that you want to answer). model enable you to build momentum and develop sufficient
Where can project data analytics be deployed to deliver • Then develop a plan to close the gap. integrated solutions to achieve critical mass?
maximum benefit? What are the current performance issues • Also develop a data dictionary to help drive commonality. Where will data be saved? How is it permissioned? Do you
within your project delivery organisation? Are there any areas have a centralised data lake or common data environment?
that have proven particularly stubborn? Your sponsor or It can take up to 10 years to gather sufficient data to support How will GDPR issues be considered? How will you buy
senior responsible owner (SRO) and senior project leadership high end use cases so the quicker that you start this journey compute to train your models? This rapidly gets into the
teams will have a clear line of sight to current areas of the sooner you’ll realise the benefits. Imagine a world where realms of your IT department. But you’ll need to understand
concern. Gaining early visibility of issues before they catch fire main contractor A has a deep and comprehensive repository how it all fits together so you can operate within it. Some
is a major driver for many. There will be an element of chicken of data and contractor B has a more ad hoc approach. organisational IT departments are enablers to the business,
and egg; if you don’t know what can be done, you’re likely to Contractor A will be able to deliver machine learning based working collaboratively to put the infrastructure in place
artificially constrain your horizons. It’s important to develop an insights two-10 years before contractor B; this could seriously to enable you to deliver your vision. Others will see their
understanding of the art of the possible. impede contractor B’s competitiveness. role more in terms of governance and will be focused
on protecting the business. You’ll need to gauge how to
There is a correlation between the opportunities/problems approach them and set out a plan to win them over. From
Data health-check or audit that we would like answered, the magnitude of these experience, this could be the biggest challenge for most
opportunities/problems (sizing the benefit and disbenefit), the deployments.
As momentum builds, you can spend more time exploring potential solutions and the availability/quality of data. This
what data you have, how good it is, what you can do with it helps to prioritise which problems to solve first. The Project
and what is missing: Data Analytics Task Force solutions development workstream Skills, roles and accountability
is developing a model to do this at a sector level, allowing
• A positive step forward is to document what data you have organisations to pool resources to solve common problems. The vision will only become a reality if someone is responsible
and where it’s stored. for it, with board level sponsorship. The best approach we
• Then assess its quality. Initially you may choose to do this have seen is when one person is responsible for developing
manually on a sample basis. As your skills develop, you’ll Data architecture and engineering the strategy, plan and implementation; including securing
introduce data science solutions to automate this process. resources, unblocking issues with IT and identifying advocates
• Then perform a gap analysis between the data that you What is your data ecosystem and how does it all fit together? who have a passion for driving the change.
Does your licencing model encourage small-scale roll outs or

13 Getting started in project data

 Contents

Taking a strategic approach

There will be a spectrum of approaches to skills and achieve a systematic application of data analytics,
development from e-learning to bootcamps and they will move further up the capability spectrum towards
apprenticeships. But project data analytics is very predictive and prescriptive analytics. Project professionals
different; our immediate challenges aren’t in the will develop superpowers to predict the future, pre-
deployment of new tools. It’s understanding how to shape empt risk and variance; they will outperform their peers,
implementation when data pipelines are fragmented, unlocking a new cadre of professional.
data is messy, schedules are emergent, risks are heavily
interdependent, particularly in highly connected projects Gareth Parkes, as Sir Robert McAlpine’s Head of Data,
with multiple suppliers. Success will require us to work fulfils such a role and has board-level support. He has
collegiately as part of a broader ecosystem. The Project been an ambassador for change, taking a key role in the
Data Analytics Task Force has a workstream, led by Sue development of the project data analytics community
Simmonite, project management and control capability as a means of raising awareness and inspiring people.
director at BAE Systems, focused on the skills challenges He shapes McAlpine’s strategy, provides the glue to
that the profession needs to grapple with. pull resources together from across the business, and
works closely with IT to address challenges with dev ops,
Professional bodies are well placed to explore future data access and licensing. He is also helping to develop
project data and analytics skills needs and working with capability within the supply chain, ensuring that the
industry and academia this could be an area for future realisation of the vision isn’t constrained by organisational
research boundaries.

Gareth has also identified a pipeline of candidates for the

Making project data analytics business as usual project data analytics apprenticeship, upskilling ~5 project
for projects professionals per cohort and enabling them to transform
their role from the inside One example is Yvanna, an
A future underpinned by project data analytics is no longer office manager who has evolved her role considerably
optional; it’s an inevitable future that will help to transform because of her new found skills, by automating processes,
how we deliver projects. The core skills and capabilities detecting documentation errors, and extracting insights.
will rapidly become business as usual, freeing us from By the end of 2021, he will have upskilled more than 30
the burden of repetitive work and enabling us to move people, a major and highly motivated force for positive
up the value chain. The challenge for your organisation is change.
how to make this vision a reality. As organisations mature

14 Getting started in project data

 Contents

What tools can help?

Overview Experience from more advanced organisations indicates “We wanted to create a single set of central reports that
that without an overarching strategy, dashboards will everybody had access to and could trust - sponsors,
There are so many different tools available, many rapidly proliferate, which become difficult to maintain. clients, project, delivery teams, audit teams. We also
providing overlapping services. Some are free, or are Mature organisations are beginning to move away from wanted to make curated datasets available so that we
already included in your organisation’s licenses. Some ‘death by dashboards’ towards role-based insights, could empower our colleagues to develop their own
can be cheap and quick to deploy, and some can be with adaptive content based on criticality. If specific dashboards and reports that meet their needs. We link
expensive and take time to deploy at scale. The key is to data exceeds defined parameters (typically set by the the outputs of our reports into the decision-making
remember they are tools to help you do a job. organisation’s PMO) it will be flagged, reported upon and process. This is important to show people that the effort
insights will be inferred, rather than providing hundreds of they’re putting into providing the data is resulting in
charts all requiring interpretation. their concerns being addressed at the right level of
Automated dashboards the organisation. Be flexible, be prepared to learn new
Simply put, project personnel can be guided as to what tools, be professionally curious, and look around for
This tends to be the place where most organisations start. they should action via automated prompts, rather than inspiration. When we developed our reports, we looked
The barriers to entry are low and the returns on investment searching through multiple dashboards and figuring out at other people’s reports, but we also looked at websites
are rapid. There’s plenty of online content to enable what their actions are which is not only time consuming and apps that people regularly use, so the interface is
you to get started and make the transition away from but can be a barrier to people adopting dashboards. The familiar.“
presentations, documents, and spreadsheets. more serious or urgent the prompts are, the higher up the
project organisation actions can flow resulting in the right Milla Mazilu
In larger projects, with lengthy reporting cycles, the report action being undertaken by the right person at the right Network Rail
can sometimes be up to two months out of date. The time. What this all means is that PMO’s have built in levels
project leadership team then lose confidence in what of assurance and ensuring the weaker members of the
they’re being told. Dashboards provide an opportunity project team better understand the expectations set of
to move towards real time reporting. But this requires a them.
move towards more advanced data engineering and
data pipelines; the end-to-end process from ingesting, • Start with simple dashboards
integrating, cleaning, processing, and preparing data. • Then drill downs
It’s likely to require the use of Python and Data Analysis • Then role-based dashboards
Expressions (DAX) which help with calculations, data • Then integrated data pipelines
integration and error checking. • Then Python and DAX
• Then automated triggers to individuals to prompt action

15 Getting started in project data

 Contents

What tools can help?

Automation Getting started with automation:

Your IT department will also have a key role to play to
The next step is to explore how automation can be used • Start with something like UI Path or Power Automate. ensure that they keep control of the implications of
to remove the human from the loop of data ingestion and Understand the principles and get a feel for what it can licencing, cloud compute etc. Ticking the wrong box or
processing. This could include everything from timesheets do. leaving computer running over a weekend can rack up
through to invoice processes. By doing this we release • Experiment with the community editions or free trials. hundreds of dollars.
capability for higher value-added activities. Note that the IT and governance challenges for some
organisations may be considerable.
There are a variety of vendors including UI Path, Blue Prism • Deploy simple solutions.
and Microsoft Power Automate. All of which begin with • Then investigate how to scale, including addressing
low code solutions, providing a low barrier to entry. This licencing requirements.
enables users to get going quickly to solve immediate
automation needs, tackling low hanging fruit. The short-
term benefits can be considerable. This could be a App development
simple workflow, or it could help to integrate disparate
corporate tools that don’t speak to each other. The authors With low code solutions the barriers to entry for app
have used automation for everything from onboarding developers are largely removed. It’s easy to develop
apprentices, preparation of contract documentation an app to record progress, trackhealth and safety
through to automated naming of BIM files in accordance observations, read invoices, track risks or similar.
with a defined standard. The capability becomes more
powerful through ‘hyperautomation’ which combines Although an app is relatively easy to develop, it can take
the workflow type automation with machine learning to time to integrate the app within the data ecosystem,
overcome some of the data structure constraints with covering everything from database integration to role-
robotic process automation. The use cases are vast. based permissions. It requires a degree of governance.
We would advocate the deployment of apps locally within
But as the level of ambition increases, the need for a project team using ‘innovation at the edge’ principles,
governance and training also grows. Some use cases but as soon as it begins to scale across the organisation
will require the processing of personal data, with GDPR or into the supply chain, it requires more careful
considerations. Others will integrate data from supply consideration.
chain systems, requiring error checking and automation
resilience, particularly where payments are involved. It
requires a higher level of skill to build code that is resilient
to changes in the location or definition of data.

16 Getting started in project data

 Contents

What tools can help?

Python Graph databases of tools will work together. There is also a vast array of niche
tools. For example, nPlan has developed a capability to
Python is a programming language that provides modules We’re starting to see a number of organisations deploy challenge schedule envelopes when the project is sanctioned;
and libraries to process and interpret data. It doesn’t take graph databases. These are well suited to a project delivery a work package level version of reference class forecasting.
long to learn, it’s free and there are vast online resources environment where there’s a high level of connectivity Nodes and Links are focusing on resolving the complexity
which can be adapted to suit your particular use case. But it between work packages, functions, risks etc. It’s now possible inherent within megaprojects. There are many others such as
helps tremendously if you understand how the code works to model these connections and leverage the database to run ones for stakeholder and team sentiment analysis or ones for
before cut and pasting it into your own model. Python opens complex queries and deploy machine learning algorithms. project assurance insights.
up vast model libraries to provide access to data pipelines, Work is in train to develop project-based ontologies, which
sentiment analysis and machine learning. capture concepts and categories in project delivery that
shows their properties and the relations between them. The
There are alternatives to Python such as R, which helps the Project Data Analytics Task Force are building, differentiating
creation of bespoke dashboards using Shiny. Solutions such between the built asset (the digital twin) and how the asset
as Alteryx offer a drag and drop functionality, removing was built (everything from contracts, risk, schedules and
some of the learning overhead with solutions such as Python. resources through to logistics and carbon).
Microsoft also offers drag and drop solutions via its Power
Platform. Your personal choice of programming language
will probably be constrained by whichever language your Project management tools
colleagues use, so that you can share and develop code.
There are a number of software vendors who offer data
Have a conversation with your IT department and understand driven solutions within their project management tool. Some
what options are available to you. Then explore the art of the offer integrated platforms, such as Procure, with the intent of
possible. Understand what can be done and how challenging providing a one stop shop for project delivery solutions. But
it is. Then start small, experiment and build outwards. If you’re organisations will need to wrestle with the overlap between
unconstrained, then learn by doing. Develop solutions and use corporate systems, their standard technology (e.g. Microsoft)
online resources to stack and new platforms. The increasing availability of
support you. APIs provide a facility for users to overcome some of these
integration challenges, allowing the integration of disparate
data sets; but you’ll need a clear strategy for how your stack

17 Getting started in project data

 Contents

Frequently asked questions

Q Q Q
Will automation take Is data analytics expensive to Do we need advanced coding skills to
over my job? deploy? deploy project data analytics?

A A A
Our Projecting the Future: The adaptive That depends on how far and wide you wish to deploy Not to start with. Tools such as the PowerPlatform
professional report (APM, 2020), identified that it and what issues/opportunities you’re seeking to offer a wealth of capabilities which are promoted
in an era of unprecedented technological, social address. Getting started may cost you nothing more as low code solutions. As you gain experience
and environmental change, the project profession than some of your own time as you may already have and gain confidence to tackle more sophisticated
will be the profession at the heart of creating and the tools you need. For example, most organisations requirements you will have gained a good
delivering change. To deliver change successfully, already use Microsoft products, so it makes sense to understanding of the coding capabilities needed
project management’s future is as the adaptive have a look at what’s provided within your current and how to get it.
profession. In order to do this, project professionals licenses, but there are a vast array of alternatives,
will need to respond to shifting contexts, continuous some of which are open-source.
learning continually, and develop new skills and
competencies such as data analytics.

Gartner predicts that 80% of today’s project

Q Q
management tasks will be eliminated by 2030, Can project data analytics works on Can data analytics be introduced for
but new ones will emerge. If we no longer have the
transformation programmes? in-flight projects or only for those being
burden of such tasks there is capacity to work on
more value-adding tasks and therefore roles that planned?

A A
we enjoy more. These are roles that have yet to Data analytics can be used on any type of project/ The short answer is both, but the approach may
materialise, but will help us to unlock a crystal ball programme. The reason there are lots of examples differ. A challenge with in-flight projects may be a
to provide predictive insights to shape decision coming from capital programmes, is because of the lack of data standards across the organisations
making. Automation may impact your job, but if you existing level of data maturity through things such involved or a lack of understanding of where
adapt and upskill new opportunities will emerge. as Digital Twins, Building Information Modelling (BIM) source data sits. But it doesn’t take long to find and
and project controls tools, so they have been early cleanse data using analytical support tools.
adopters. But we’re seeing data analytics being
used in helping transformation programmes gain
a better understanding of stakeholder needs and
relationships.

18 Getting started in project data

 Contents

Frequently asked questions

Q
How is APM addressing the impacts of data • working with our Specific Interest Groups (SIGs) and branches
analytics on the project profession? to discuss the PDA developments including the role of data
analytics in specialist areas. For example, the Planning
Monitoring and Controls (PMC) SIG have created a PDA working

A
We have undertaken a wide variety of activity to raise group, has set up a PDA sub-group which aims to share
awareness and understanding of project data analytics data ideas and insights from across organisations. For more
and its impact on the profession by: information on this please contact Milla at [email protected].

• convening a Data Advisory Group which brings

together a broad range of likeminded organisations
whilst aiming to improve ‘data literacy’ enabling
individuals and organisations to understand how they
might be able to make better use of data;
• funding and undertaking research in relation to project
data and analytics. This has included the creation
of a PDA Pathfinder report, Project Data Analytics:
the state of the art and science, led by the WMG and
partners which provides a concise overview of project
data analytics. Our Dynamic conditions for project
success 7 highlighted the importance of project data
and technology identifying this as one of its nine
dynamic conditions for organisational success.
• producing a broad range of resources around the topic
including blogs, podcasts, events and webinars and a
data analysis and storytelling for success e-learning
module available for our members. These resources
can be accessed at apm.org.uk/resources/what-
is-project-management/what-is-project-data-
analytics/;
• collaborating with partners such as supporting the
development of a project data maturity model being
delivered by WMG at the University of Warwick;

19 Getting started in project data