Big Data 2019 Speaker Interview: Dr Clair Sullivan

Big Data 2019 takes place at the ICC Waterfront in Belfast next Thursday and the conference will play host to some amazing speakers. On Friday I had the opportunity to interview one of their most exciting speakers, Dr Clair Sullivan of GitHub. Dr Sullivan is a machine learning engineer at GitHub working on large graphs of users and code and is travelling from Colorado to speak at Big Data. The title of Clair’s talk will be Data Science; Revolution to Evolution.

Data Science; Revolution to Evolution

Clair’s talk will centre on the history and roots of data science and how it informs our future. Now that the term ‘big data’ has been in common usage for maybe ten years, we have come to realise that not everything involving data science is revolutionary. With Clair, the Big Data audience will take a look back at what we have accomplished through data science and what it’s done for us.

Clair pointed out that there is a perception that great achievements only come through revolution but small evolutionary steps should be equally embraced.

Did she ever imagine that she would be a Data Scientist or Machine Learning Engineer?

I asked Clair how she came to be in this area of science and she confessed that it has been a strange journey. As a child, she wanted to be a cardiac surgeon but then she discovered that she liked physics. After attaining degrees in physics and astrophysics, she realised there were few jobs in those specific areas and wound up in Nuclear Engineering.

..and then came 9-11

9-11 happened a year before Clair received her PhD doctorate in nuclear engineering. Clair realised that some of the research she was working on could have helped detect terrorists and her professor agreed. After 9-11, the US government were looking for people working in radiation detection and Clair got involved in analysing data coming off those detectors, working in nuclear emergency response researching spectroscopic data. The skills she developed here are the same skills used by data science today.

After a while, Clair wanted to spend time teaching and so she worked for 5 years as a professor at the University of Illinois in the Department of Nuclear, Plasma, and Radiological Engineering while researching the use of machine learning to analyse the data from large sensor networks. Her students moved on to Silicon Valley and then she got a job as a data scientist with GitHub where she now focuses on Machine Learning while maintaining adjunct assistant professor status at the University of Illinois.

Explaining the terms Machine Learning and Spectroscopic Data

Machine Learning is the art of taking data from different sources and writing math to make sense of it, and perhaps to identify and predict trends. This is the tech behind those apps that explain photographs for those with visual impairment. 

Spectroscopic Data is the information that can be gleaned from what is not visible to the naked eye, for instance, infrared, ultraviolet or gamma rays of light. Clair has been involved with the use of spectroscopic data to measure radiation through methods such as the identification of Caesium 137, a common contaminant of radiation.

Using Data Science for Good

Data Science is, of course, an incredibly broad term encompassing everything from the Cambridge Analytica scandal to medical advances and using data to help prepare for catastrophic events. While some of the biggest usages of data science has been for marketing, the greatest advances have come through the use of machine learning for social good, for instance, to augment what radiologists see when they look at a mammogram. Data protection regulations in Europe such as GDPR, do not exist in the USA, or at least there is no single principal data protection legislation there, and this has resulted in US tech companies scrambling to be compliant with European markets.

Our discussion turned to ethics in data science, and one point that Clair brought up was the realisation that there is a lot of bias towards white males in facial recognition models which leads to far lesser accuracy in the recognition of say women of colour. She said that it is part of the data scientists job to take data and drive good uses for it, while always being aware of where the biases are.

In her own side company, La Neige Analytics, a company which was borne out of a bootcamp and her own passion for skiing, data science is utilised to look at satellite imagery to determine the amount of water in the mountains. This is the kind of technology that can also be utilised in predicting likely flood points, etc. Interestingly, Clair told me that a large part of the learning with La Neige centred around communication between data scientists and non-science management, and how improving this interface better informs business society who benefit from the many answers that science provides.

Mentoring

Clair is an advocate of mentoring and currently has a mentor matched by GitHub. Surprisingly, her mentor was someone in Australia who has no college degree, but who she nonetheless finds brilliant and who mentors her in the areas of communications and the conversion from an academic to a more industrial existence. She also proudly mentioned a student that she recently mentored who now works for a very well-known ride-sharing company.

What living person does she most admire?

Clair said she was torn between two living people who she admires greatly and they are Barrack Obama and the Dalai Lama. Her reason is that while both have been the targets of some nasty slurs and attacks, both have maintained a positive public face. While she believes they would both have every reason to hate the world, they don’t but are instead encouraging, positive and hopeful. It is this positivity and goodwill towards others in the face of adversity that she admires most.

What is the best advice she ever received?

Clair once had an Astrophysics undergrad mentor who was friends with Stephen Hawking! (!!!) Her professor had many funny stories about Hawking who was naturally revered in the scientific community (and beyond). This mentor told Clair to remember that the smartest of the smart are also usually the nicest people too because they have nothing to prove – they don’t need to be worshipped, the only thing they need is what we all need, and that is friendship and compassion.

Belfast

Dr Sullivan has never been to any part of the island of Ireland before and told me she is really looking forward to visiting Belfast. Her husband comes from an Irish family and she is clearly up-to-date on political issues here such as Brexit. While in Belfast, she hopes to visit the Titanic Museum and she tells me she is also a Game of Thrones fan so hopefully, she’ll also have time to visit some GOT sites.

Speaking with Clair Sullivan through the medium of Skype was an honour for me and I greatly look forward to meeting her in person and to hearing her talk when I attend Big Data next Thursday.

Read more about Big Data 2019 in this preview article: Big Data Returns to Belfast

***