Meetup Data Science in Industry
On Thursday February 23, 2017, VORtech hosted our first Meetup on Data Science: “Data Science in Industry”, in collaboration with the Meetup group Data Science Rotterdam. The subject was strongly related to one of our key specialisms: data science. There was food, three talks about data science in industry on three very different topics, and an opportunity to network afterwards.
At VORtech we are fond of the Meetup platform and what it stands for: to come together with like-minded people, freely share information and learn from one another. Our colleagues regularly attend Meetups and VORtech has hosted several Meetups in the past, mostly on topics related to High Performance Computing. And this time, the meetup is about Data Science.
Parameter calibration and data-assimilation for indoor climate modelling using OpenDA (Werner Kramer, VORtech)
The first talk was presented by Werner Kramer (Scientific Software Engineer at VORtech). He spoke about “Parameter calibration and data assimilation for indoor climate modelling using OpenDA”. Indoor climate modelling has a number of uncertain variables: forcings and physical parameters are not exactly known, the airflow field is turbulent, and the numerical model is not an exact representation of reality. VORtech uses OpenDA to combine the numerical airflow model with observational data to improve the model results. VORtech applies calibration methods to historical data to obtain better parameter estimates and data assimilation to further improve forecasts.
Advanced analytics at Shell (Winfried Theis, Shell)
After Werner’s talk, Winfried Theis (Lead Data Science in the Advanced Analytics CoE at Shell) continued with his talk on Advanced Analytics at Shell. In the energy sector, large volumes of data are already available for many decades. What is changing in recent years is that these large volumes require real-time analysis and are coming from different sources. This talk gave an overview of the organization behind this transformation and three use-cases from different areas where data science and statistics have contributed to the success of our business.
Data science and noisy social media data (Hannah Tops, Coosto)
Finally, the third speaker was Hannah Tops (Data Scientist at Coosto). She spoke about Data science and noisy social media data. Coosto is a social media monitoring tool and being a Data Scientist at Coosto means we get to work with billions of messages. The majority of those messages come from social media platforms, which means they are short, misspelled and without context. In this presentation, she showed the difference between a model trained with “nasty” social media and one with cleaner data. She furthermore showed how we build a model that recognizes a “buy intent” in messages and, lastly, how we handle feedback from and communicate with the business people who are using our models.
This event was organized and sponsored by VORtech and attendance was therefore completely free. Due to capacity constraints, we could only accept the first 60 RSVPs.