How to Better Apply New Digital Technology in Science
Despite the digital revolution, scientists have been using the same workflow for data management and collaboration for the last forty years.
Platforms for science tend to replicate paper-based documentation and folder-based organization on the computer screen. As a result, many scientists see no need to exchange paper lab books for a digital platform. The scientific process, from data acquisition to publication, could be greatly improved by recent advances in network theory, artificial intelligence and data mining.
Below I address four aspects that could improve the quality of digital platforms for scientists.
1. A Better Perception of Scientist Profiles
In the last years, the pharmaceutical and healthcare sectors have polarized the perceptions of developers and investors. The work environment of these sectors has become the reference point for the software industry. If we look beyond industrial research, a much more colorful picture of the scientific environment emerges.
While some scientists use commercial instruments in routine labs, others build their own measurement tools. Some scientists possess the technical capacity to write programs, while others have more limited computers skills. Scientists may work in highly restrictive environments or in flexible project structures with significant autonomy. Some professors and group leaders carry out their research projects alone, while many others work with intense international collaboration.
Successful research teams in institutes and universities are made up of a diverse group of scientists with varying profiles. New software concepts should create interfaces between scientists with complementary skills.
2. Innovations in Processes Management
The production of scientific knowledge cannot be compared with a car factory, where each task or procedure is rigidly serialized. It is rather an iterative process, where the scientist gets closer and closer to the sought result. Attempting to frame the scientific process in Gantt or PERT-like charts is an inadequate approach.
Would it not be more useful to manage iterative processes by borrowing Agile concepts from the software industry? Scientists have already begun to use tools like team messaging and Kanban boards, which today are popular among software developers. Current communication and network technologies enable new concepts of management that were not previously possible. Modern and dynamic task management tools for scientists would be greatly appreciated.
3. Data Analysis Beyond Statistics
Thanks to the popularity of e-business and stocks analytics, there is a large selection of software with statistics tools already available. However, statistics calculations are just a small fraction of the data analysis performed by many scientists. There is a lack of software options for general data analysis.
The currently available options are expensive and often demand programming skills from the user. Few researchers are able to develop their own software solutions to target a specific problem. In worst-case scenarios, scientists will end up juggling multiple spreadsheets to complete a research task.
The current software deficits leave ample opportunities for development in this field. Artificial intelligence algorithms, for instance, could empower platforms with customizable and adaptive analysis environments. Uniting analysis know-how and experimental data would help to spread scientific knowledge and reduce the dependence on experts in the future.
4. Sharing and Collaboration
Resistance to cloud solutions remains an obstacle for many scientists who could benefit from online data-sharing platforms. Part of this resistance is based on the false premise that cloud collaborative platforms must be inherently connected to open data initiatives.
The free sharing of information is an ideal defended by all scientists. However, it should be noted that scientists are not willing to share their data until the result of the research is published. Science is a highly competitive field and scientists are invested in protecting their research until publication.
A significant improvement would be the creation of more cloud platforms for sharing and working on data before publication. Access to this data should be restricted to collaborators. Such platforms would strengthen collaborations and allow data investigation to occur more dynamically. Scientists would be able to reach conclusions and publish their results more quickly and the data would become more promptly available to the entire community.
Summary
Software developers need to establish closer contact with scientists from different fields. Platforms should be designed to take into account the actual needs of the scientists on the lab floor. Investors need to take note of the strong demand for data management solutions in areas other than the life sciences, including, for example, the material sciences. We mustn't be afraid to try out creative solutions. Let’s bring all scientists into the 21st century.
About the Author: Dr. Carlos Viol Barbosa worked for twelve years in material science and nanotechnology. He is CEO and co-founder of ScienceDesk.net.
