Supply Chains Analytics: A Rich Smorgasbord For Data Scientists To Feast On!

On July 26th, the 38 DatSci Awards finalists will be going head to head for a full day of presentations at Croke Park, Dublin and giving our Judging Panel compelling reasons and insights as to why they deserve to win a 2018 DatSci Awards Trophy.

Mick Phelan, Business Technology Leader in Data Science in Johnson & Johnson is a finalist in the Data Scientist of the Year Category powered by Deloitte. Check out what Mick has to say about Supply Chain Analytics. 

Supply Chains are complex, frustrating and fascinating in equal measure. While they are not at the top of everybody’s list when it comes to data science jobs; they are rich with data science opportunities, including:

  • predicting customer behaviour & lead times (e.g. 24hr delivery for a product that takes 12 weeks to produce)
  • predicting the impact of in-store & online sales campaigns
  • forecasting product quantities 4-6 months before they are ordered (goods manufactured in Asia & shipped via ocean to US or vice-versa)
  • mitigating the Bullwhip Effect
  • people and machine capacity planning optimisation
  • predictive asset maintenance (PdM) using deep learning
  • production scheduling optimisation
  • quality anomaly detection

The list goes on, but hopefully, this gives a small flavour for common areas data science can add value across the supply chain.

Traditionally when we think of supply chains, we think manufacturing and labour intensive factories. According to the IDA, manufacturing accounts for more than 36% of Irelands GDP, employing over 11% of the workforce directly, and 22% across all skills levels when indirect employment is taken into account. The face of manufacturing is changing rapidly and the IDA is keen to establish Ireland as a global hub for advanced manufacturing technologies including automation, IoT, 3D printing, VR & AR and data science.

We have a rich history in Ireland for global market-leading supply chain & manufacturing operations, pushing the envelope within multi-nationals to win more business and be the best performers in the supply chain. With this ethos, we can establish Ireland as a global hub for data science across multi-nationals and our indigenous organisations. We have the companies, a small, well-connected country, an exponential rise in data science / AI courses across the Irish academic sector in the last 3 years, and most importantly, the people to make it happen.

Estimates of the economic impact of digital manufacturing vary widely. A widely-cited figure is an estimated US$100 trillion of value that digitalisation could create to global industry and society over the next decade. But, what are we going to do with all this digital data, streaming 24×7 into our data lakes? The healthcare supply chain is changing rapidly, mirroring that of the electronics sector over a decade ago, and technology is playing a more prominent role. Personalised healthcare will change how we consume products and services, pushing supply chains to be leaner, more agile and responsive while lowering costs of production.

To quote Peter Drucker: ‘culture eats strategy for breakfast’. This is often the case in supply chains when it comes to data science, stymying its adoption. Promoting analytics in manufacturing requires a multi-faceted approach. Working in partnership with our DePuy Synthes Operations partners, the Data Science Group developed the “Lean Analytics” concept.

The essence of Lean Analytics recognises the similarities between Lean and analytics adoption in an organisation. Both domains provide practitioners with a set of tools and techniques for tackling problems, with a clear emphasis on understanding the problem statement; without this success is unachievable. Lean is widely promoted across the manufacturing sector, with most organisations boasting operational excellence groups; encouraging a similar adoption to data science breaks down barriers.

Once the problem statement is defined (e.g. using A3 problem-solving techniques), relevant business SMEs and data sources are identified, and an “app” based approach is employed to answer the question. If you think about apps on your mobile phone, they don’t require a manual; generally, they perform a single task and do it very well (depending on the app!) in an intuitive manner. Currently, there are 3 apps on the Lean Analytics platform (all web-based):

  1. Production workflow
  2. Capacity planning
  3. Production attainment

Each app employs existing Excel reports for use by one or more departments in the manufacturing process. The user copies the reports to a folder and the app does the rest: extracting, cleaning, transforming and blending data into information packets. Each app has a different advanced analytics model: probability tree prediction, mixed integer programming and lead time prediction. All apps have a Qlik Sense visualisation enabling the user to interrogate the results from different angles (or features if you’re a data scientist). This ecosystem enables visibility (answering the underlying questions) across the manufacturing process and is portable to other sites in the network.

The rapid pace of evolution across supply chains to meet customer expectations is set to accelerate as the next generation of consumers enters the marketplace. Demand for near-instant delivery of products and services is on the horizon, placing unprecedented pressures on supply chains. The emerging technologies of today (IoT, AR, VR etc.) will be superseded by the next wave of technology we have yet to develop. To prepare for this exciting future, we need to invest in our primary and seconds schools, ensuring our children are equipped with the necessary skills to fuel this innovation.

The DePuy Synthes Data Science Group developed a Lego robotics model to play noughts and crosses (Tic Tac Toe) optimally; employing the Minimax algorithm from Game Theory, all coded in Python in the guise of a Lego Technic model. This robot (OXO) is a demonstration of the intersection of various disciplines (mathematics, computer science & engineering) aimed at stimulating younger students’ interest in STEM subjects.

Following the interest in OXO, we piloted a Lego robotics course with 4th class students from Summercove National School, Kinsale, Co. Cork. Over 10 weeks, students (10-11 yrs) learned to adapt, build and program robots; developing their engineering, design, technology, computer science and mathematics skills, complementing the school curriculum. Taking concepts learned in the classroom (distance, degrees, rotations, logic, etc); students worked with motors & sensors, programming their robots to carry out different tasks; developing workflow skills and learning about logic, loops, switches and more! Following a Lego “Fort 2 Fort” Challenge, all students received a certificate of completion.

After the 10 weeks with 4th class from Summercove, students exceeded all expectations, adapting to the concepts, logic and interfaces with ease. With the abilities and expectations these students have in primary school, they will propel demand for services and products at a scale yet to be witnessed; pushing the boundaries of what our supply chains can achieve whilst providing the much-needed solutions and leadership to ensure these supply chains deliver to meet the needs of tomorrow.

Data science is an integral part of the changing supply chain landscape, it should no longer be considered a niche discipline but as integrated as planning, operations or engineering. We are generating an exponential volume of data (estimated to double in the next 2 years) and are struggling to keep pace. The next decade will see unprecedented changes across healthcare supply chains; evolving existing fields of study and embracing new ones. Data will be the core of this change and Ireland Inc. is well-placed to be a driving force behind the next generation of supply chains.

Mick Phelan is one of the 38 finalists in the 2018 DatSci Awards.  The Awards Ceremony is an opportunity to connect with some of the country’s finest Data Scientists, be sure to save your seat!

By |2018-07-24T16:53:34+00:00Jul 24th, 2018|

Leave A Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.

DON’T MISS OUT!
DatSci 2019 Updates
Be the first to get latest updates and exclusive content straight to your email inbox.
close-link