What is Data Science? We asked our Data Platform Director, James Frost
What is Data Science? We asked our Data Platform Director, James Frost
We are very proud to say Quorum’s Data Platform Director, James Frost recently received news that he had obtained a distinction for his MSc in Data Science.
But what was it all about, and why did he do it? We decided to sit down and ask him about it.
What Is Data Science?
Data science is quite simply applying scientific method to the analysis of data. It is an extremely wide (and often misunderstood) field, incorporating statistics, data warehousing, programming (Python, R and SQL are the most common data analysis languages), big data techniques and machine learning.
The important thing to understand is its not about the technologies you use, but the care with which you do your analysis so you can say with a high degree of certainty that your results are valid and should be acted on. There are plenty of examples in the media of poor data science where people jump to conclusions based on limited information and dubious analysis.
Why Did You Decide To Do This MSc?
I started the MSc two years ago, as I could see this was an area that our clients were becoming increasingly interested in and I saw this as an area where we could really add value to our client’s businesses. Quorum already have a lot of experience in traditional database and data warehousing / reporting techniques and I felt this was a great foundation to then be able to apply more sophisticated analysis techniques to extract even more value out of this business data.
How Was The Course?
The MSc at Dundee University has transformed the way I view data – the course was mostly completed via distance learning but with 2 intensive weeks per year at Dundee, led by Professor Mark Whitehorn. Subjects covered in the course included Hadoop, Spark, NoSQL (Mongo, Graph, Cassandra), Relational Theory, Data Warehousing and Machine Learning. There were also some really interesting projects including a computer vision assignment (recognising Formula1 cars in video footage), a Zombie Apocalypse “Monte-Carlo simulation”, a data mining project, and papers on Hadoop and deep learning frameworks. My final project was to build a neural network that learned to play backgammon via reinforcement learning (based on a similar approach AlphaGo Zero).
Now You Know More About It, What Misconceptions Do You See About The Data Science Industry?
People enter data science from a wide range of backgrounds so the term “data scientist” is often confusing, and companies looking for a data scientist need to think carefully about the blend of skills they are actually looking for – for example some will need a statistician to do complex mathematical analysis, others a business intelligence expert, whilst some might need an expert in machine learning to extract previously unknown facts from a companies twitter feed. Good data scientists should have a wide range of understanding to know the techniques you could apply to problems such as these, but will not be able to be experts in all of them.
Companies also often want to jump into doing advanced techniques such as machine learning, without wanting to spend time on data quality issues – this just leads to bad analysis and often confusing or conflicting information being produced by the resultant models. Advanced data science techniques need to be built on top of solid foundations, such as having a robust data warehouse in place.
What Is The Future Of Data Science?
One thing that has been staggering is the explosion of deep neural network techniques over the last 3-4 years, with companies such as Microsoft, Google and Facebook leading the field but also sharing their techniques, data and frameworks with the wider community. This has been driven by the rise of publicly available big data sets for training, frameworks such as Tensorflow and Keras, and a huge increase in compute capabilities (in particular graphics processing units).
Whilst General Artificial Intelligence is still a long way off, Narrow AI has surpassed human intelligence in a number of fields including boardgames (Go, Chess, Shogi), computer games such as StarCraft, language translation, image recognition, and even the ability to land re-usable rockets for SpaceX. Areas such as self driving cars, medical analysis and legal document creation are high on the agenda for the next wave of changes, which all have the potential to bring about huge improvements to society.
You can find out more about Quorum’s Data Solutions HERE
We are very proud to say Quorum’s Data Platform Director, James Frost recently received news that he had obtained a distinction for his MSc in Data Science.
But what was it all about, and why did he do it? We decided to sit down and ask him about it.
What Is Data Science?
Data science is quite simply applying scientific method to the analysis of data. It is an extremely wide (and often misunderstood) field, incorporating statistics, data warehousing, programming (Python, R and SQL are the most common data analysis languages), big data techniques and machine learning.
The important thing to understand is its not about the technologies you use, but the care with which you do your analysis so you can say with a high degree of certainty that your results are valid and should be acted on. There are plenty of examples in the media of poor data science where people jump to conclusions based on limited information and dubious analysis.
Why Did You Decide To Do This MSc?
I started the MSc two years ago, as I could see this was an area that our clients were becoming increasingly interested in and I saw this as an area where we could really add value to our client’s businesses. Quorum already have a lot of experience in traditional database and data warehousing / reporting techniques and I felt this was a great foundation to then be able to apply more sophisticated analysis techniques to extract even more value out of this business data.
How Was The Course?
The MSc at Dundee University has transformed the way I view data – the course was mostly completed via distance learning but with 2 intensive weeks per year at Dundee, led by Professor Mark Whitehorn. Subjects covered in the course included Hadoop, Spark, NoSQL (Mongo, Graph, Cassandra), Relational Theory, Data Warehousing and Machine Learning. There were also some really interesting projects including a computer vision assignment (recognising Formula1 cars in video footage), a Zombie Apocalypse “Monte-Carlo simulation”, a data mining project, and papers on Hadoop and deep learning frameworks. My final project was to build a neural network that learned to play backgammon via reinforcement learning (based on a similar approach AlphaGo Zero).
Now You Know More About It, What Misconceptions Do You See About The Data Science Industry?
People enter data science from a wide range of backgrounds so the term “data scientist” is often confusing, and companies looking for a data scientist need to think carefully about the blend of skills they are actually looking for – for example some will need a statistician to do complex mathematical analysis, others a business intelligence expert, whilst some might need an expert in machine learning to extract previously unknown facts from a companies twitter feed. Good data scientists should have a wide range of understanding to know the techniques you could apply to problems such as these, but will not be able to be experts in all of them.
Companies also often want to jump into doing advanced techniques such as machine learning, without wanting to spend time on data quality issues – this just leads to bad analysis and often confusing or conflicting information being produced by the resultant models. Advanced data science techniques need to be built on top of solid foundations, such as having a robust data warehouse in place.
What Is The Future Of Data Science?
One thing that has been staggering is the explosion of deep neural network techniques over the last 3-4 years, with companies such as Microsoft, Google and Facebook leading the field but also sharing their techniques, data and frameworks with the wider community. This has been driven by the rise of publicly available big data sets for training, frameworks such as Tensorflow and Keras, and a huge increase in compute capabilities (in particular graphics processing units).
Whilst General Artificial Intelligence is still a long way off, Narrow AI has surpassed human intelligence in a number of fields including boardgames (Go, Chess, Shogi), computer games such as StarCraft, language translation, image recognition, and even the ability to land re-usable rockets for SpaceX. Areas such as self driving cars, medical analysis and legal document creation are high on the agenda for the next wave of changes, which all have the potential to bring about huge improvements to society.
You can find out more about Quorum’s Data Solutions HERE
AWARDS & RECOGNITION
FOLLOW US
CONTACT INFO
Quorum Network Resources Ltd
18 Greenside Lane Edinburgh
UK EH1 3AH
Phone: +44 131 652 3954
Email: marketing@quorum.co.uk
CONTACT INFO
Quorum Network Resources Ltd
18 Greenside Lane Edinburgh
UK EH1 3AH
Phone: +44 131 652 3954
Email: marketing@quorum.co.uk
FOLLOW US
AWARDS & RECOGNITION