Every day we hear the term “data science”** with greater frequency, and its popularity doesn’t come without reason. At WillDom, we have more and more data scientists entering our community who collaborate with software developers, engineers, and other technologists, and are playing a key role in the continued success of our clients.
In this post, we’ll discuss the role as well as two data science case studies that one of our computer engineers, Verena Ojeda, has worked on, showcasing the power that these skills can have not only on business but humanity.
So why is data science so important? Data science is a field of study that extracts knowledge and meaningful insights from both structured and unstructured data. This interdisciplinary field incorporates mathematical and scientific methods, algorithms, and systems to prepare and deliver information with accuracy and credibility.
The field has been receiving great recognition from companies and businesses of all markets and sizes who are realizing the benefits of the insights to help strengthen their operations:
- More accurate and credible decision making
- Identify and act on trends in the market
- Pinpoint weaknesses/ areas for improvement
- Efficiency – accelerate analysis and learning
- Innovation (exploration of new approaches)
- And many others
Here are a couple of data science case studies that Verena has spearheaded that have provided critical knowledge and insights to help drive real change.
Case study 1: Predicting Dengue Virus Outbreaks in Paraguay*
Dengue, a mosquito-borne viral infection, is found in urban areas of tropical climates worldwide. It has a strong presence throughout South America, even being classified as an epidemic to the region. This caused Verena and her partner to focus on the country of Paraguay and the predictions of the infection’s outbreaks in the region. They hypothesized that geography, climate, and population were variables involved in the spread of the virus.
They laid out a simple overview of the process for the project:
With the assistance of the Ministry of Health, they acquired data on Dengue cases since 2013. They also received data from the Climate Directorate on climate variables for the years of the research as well as the Directorate of Surveys and Census for data on the population. This secondary data allowed them to bring greater efficiency to their process by cutting down on research time.
After taking time to standardize and anonymize (specifically the sensitive health data), Verena and her partner utilized advanced learning algorithms like random forest and decision trees to conduct the solution search. They then analyzed their results until finding an effective solution to be interpreted and implemented.
Along with answering some of their primary questions, this study also allowed them to become more aware of the roadblocks that can arise when obtaining data and also processing it.
Case study 2: Automatic detection of toxoplasmosis in Fundus Images*
Verena had the great opportunity to collaborate on a project to automate the detection of toxoplasmosis from fundus images. The team initiated the research phase by establishing an imaging protocol and collecting data. It wasn’t long before they realized how inefficient their data collection process was, which brought them to look at secondary research supplied by members of a similar project also working on fundus images.
We then began our experimental phase in which they used deep learning. They did this by training residual neural networks to identify images with lesions caused by toxoplasmosis and then categorizing them. The implementation of data science and its interdisciplinary techniques allow them to not only answer primary research questions but also develop new ones as they continue their research. Verena and the team are looking to explore new areas of deep learning techniques such as analyzing experiments regarding interpretability to reach a favorable solution for their research.
These data science case studies leverage methods and systems that allow for meaningful information to be generated, which is then turned into viable solutions with deep learning. The possibilities of research become endless as data science raises new questions and opens doors to explore new ideas with each result. Data defines the field and as we learn more, its importance continues to grow each day to have a broader impact on our society.
Hopefully, these data science case studies provided you more background behind the role and the endless possibilities within this discipline. To learn more about how our data scientists can create real change for your organization (and the world), please visit us at Willdom.com or follow us on Linkedin for more relevant content about technology trends.