Women in Data Science (WiDS) started as a conference at Stanford in November 2015. It is an initiative to showcase outstanding work done by women in the field. Now, it has evolved into a global conference, with more than 200 events worldwide. This year Maastricht University hosted its first ever Women in Data Science event, organised by the Institute of Data Science to inspire and educate data scientists, regardless of gender, and to support women in the field.

Our Digital Society Health & Well-being Professor Lisette van Gemert-Pijnen from the Persuasive Health Technology research lab at the University of Twente had the honour to give the first keynote lecture at the WIDS conference in Maastricht. Van Gemert-Pijnen’s inspiring presentation titled ‘A tech driven society, who rules the data?’ addressed the key concerns about Big Data: fear of losing control or autonomy, how to guarantee the quality of data, how to avoid data gaps (e.g., missing data on women in medical studies),  and how to address data biases that arose due to a myopic view on data collections.

In her talk, professor van Gemert-Pijnen explained why a human touch is key for responsible data science that promotes inclusiveness, equity, diversity, integrity and solidarity. Historically, data biases have been created by a lack of data collection on some groups of society (e.g., women). Unfortunately, when data are not collected decisions lack precision and their effect cannot be accurately assessed. For example, male crash-test dummies have been used in the design of cars, even though a women’s body is significantly different in at least size and muscle mass. This has greatly impacted women’s safety in cars as they are not designed around their safety needs. Data wisdom on the other hand requires the effective use of knowledge in decision making and to create value, personalized solutions are necessary. In the healthcare sector, this will be possible by combining data from wearables (e.g., pressure and temperature) with medical data (e.g., medical history, hypertension) and knowledge from psychology (e.g., how to motivate people to live healthy) and design (e.g., just in time feedback). Combining these data could be used to build an effective self-management program that helps to predict and prevent the risk of foot ulcers in patients with diabetes (e.g., an early warning system).

Van Gemert-Pijnen also discussed potential solutions for a very timely topic, namely infectious diseases.  To prevent and manage infection outbreaks we need data to understand the spread of infections. For example, to identify and predict the spread of infections in real time, we need a variety of data like movements of personnel (e.g., geospatial data), compliance to measurements (e.g., hand hygiene), epidemiological data, flight records and social media input. Interdisciplinary interpretation of these data from several disciplines, including epidemiologist, behavioural scientists and engineers is needed  to make sense of all these datapoints. In addition, it is crucial to inform local, national governments, industry and public health organizations of the findings in a timely manner. AI algorithms built with high quality data that could be developed to speedily identify what preventive measures will have the greatest effect on stopping the spread of the disease.

The other talks at the conference included:
Sofie de Broe, Centre of Statistics, Netherlands “A new model for National Statistical Institutes”
Lyana Curier, Centre of Big Data Statistics, Netherlands “On machine learning and remote sensing”
Wanting Huang, Accenture, Netherlands “How blockchain could improve the way we manage our health data”
Stavroula Mougiakakou, Bern University, Switzerland “Translating food images into nutrient information: AI for dietary assessment”
Helena Deus, Elsevier, US “Deep learning in life sciences and healthcare – stories from the trenches”
Katleen Gabriels, Maastricht University, Netherlands “Machine(s) learning morals”

Two prizes were awarded during the conference. The price for the Women in Data Science Datathon challenge, predicting patient survival based on 130.000 hospital Intensive Care Unit visits, went to a gender balanced team from UM. The Data science Research Competition price was awarded to the team investigating treatment dropout in depression.

The Women in Data Science conference was partially funded by Diversity and Inclusivity Grant of Maastricht University and the Aspasia grant from NWO.