Introduction
Recently, I had a great experience teaching a 100% remote Deep Learning course to a small group of people from one of the largest companies in Germany. The lessons I have learned as a teacher, as well as my countless previous experiences as a student, motivated me to describe my views on the key factors when it comes to choosing a Data Science course for the upskilling of your workforce.
The course
As I mentioned previously, the course dealt with the topic of Deep Learning and lasted for eight weeks. We worked eight hours a day and changed the subject every new week, including a final project where we evaluated the skills learned during the course. The main topic of the classes was neural networks, including Recurrent Neural Networks and Convolutional Neural Networks, and their practical application in a business environment.
An example of a typical class week includes:
- Theoretical classes, where the key concepts of the subject were introduced and clarified.
- Practical classes, where the programmatic implementation of the previously learned concepts was demonstrated.
- Self-study time where each student could deepen his/her knowledge and solve assignments given by the teacher.
- Support hours for questions regarding the week’s subject.
- Closing of the week to retrospectively analyze the week’s performance and review the concepts learned, as well as the materials used.
A remarkable aspect of the course was its modality, where the teacher only gives theoretical and practical classes once a week, and the rest of the time, the students work with the flipped classroom method. This way, they learn the contents at their own pace, making the interaction with the teacher complementary to their own progress. In a nutshell, this means that the teacher becomes more of a tutor, helping to guide the student, encouraging group work and constructive interaction so that students support each other and share their knowledge, making the experience beneficial for all. This modality has several advantages when implemented in such a course. Mainly, it reinforces group dynamics, which play a significant role in motivating students, both individually and as a group.
On an individual level, I could see that many students used their free time to ensure that everyone in the group understood the content learned that week. That generates a constant learning environment, in which students alternate the roles of student and teacher within the group, depending on their needs and abilities.
Personally, I felt that as a lecturer, I had more freedom to evaluate group and individual progress at all times. Perhaps it was because of the remote mode of the course, but it was much easier to have spontaneous conversations with the students to support them in their personal learning and to exchange feedback about the course. That allowed me to change the course activities and content according to the group’s daily or weekly performance and needs so that no one was left behind.
Skillset required to take a Data Science course
Learning Deep Learning requires a previous set of skills that are necessary to understand the concepts and the technical aspects of this subject. I do not want to delve too much into this topic, however, in my opinion, it is strictly necessary to have prior knowledge and experience in mathematics, statistics, programming, algorithms, and business. Not having the knowledge base to deal with topics of such complexity as Deep Learning can be highly counterproductive for students. I strongly recommend that students gradually develop a solid knowledge foundation, especially in mathematics and programming, before learning the more advanced areas of Machine Learning. Having that knowledge will make life easy for them in the future and allow them to learn faster and better.
Points to consider before choosing courses in the area of Data Science
In my opinion, before selecting a course in the area of Data Science to upskill your workforce, it is of great importance to ask yourself questions about what you want to achieve with it in the first place. In order to get the most out of such an investment, you have to find synergies between what you want to achieve as a company and what your employees want to achieve at a personal level with their training.
Define the digital maturity of the company
Doing this will tell you a lot about the status quo in terms of the skills that are already present in your company and that form the basis for the skills your company and employees will need in the future to make a positive impact on your company’s business. For example, if within your organization, all the data and reports are Excel-based, probably having employees specialized in Deep Learning will not bring any benefit to your company, since, at the moment, you do not have the necessary base to carry out Deep Learning projects. Surely in this scenario, it would be more beneficial for the company to consolidate the data in a central database and have a modern and accessible reporting through a Business Intelligence tool. Along the same lines, company culture also plays a relevant role in digital maturity. If your employees are not used to basing their decisions on data, the best analysis, predictions, and recommendations will be useless, because there is no organization-wide methodology for making decisions based on such tools. It is important to know how to evaluate where your company is now and where it intends to go in the future, taking into account the technologies and skills you already have, so that you can take steps in the right direction and see results.
Recognize potential areas of improvement in your organization
When assessing what skills a company needs, it’s a good idea to identify processes and activities that have the potential to be improved with data-driven solutions. To achieve that, it must be established whether, for example, the necessary data exists, whether it is in the quality and form that is needed and whether it is stored in an accessible place. In the best case, there is already a study of possible Data Science use cases within the company, which acts as a roadmap and provides the necessary information to know the steps and skills needed to carry out these cases.
Having this information and sharing it with your workforce before enrolling them in a course is of great value, as it will create a learning experience where they are focused on the “real” problems the company has. That will act as a sort of compass during their learning and will surely encourage them to learn and try out different things to solve those problems in the future, using their newly acquired skills. That point in my opinion, is essential if you want to ensure a win-win situation for the organization as a whole, where value is generated for everyone.
Create a Data skills map for your company
Consider the skills your workforce already has before starting the course and carefully assess their level so you can choose the right program for each of them. This way, you can avoid overwhelmed or unchallenged employees in the future. Always remember that learning new skills and being productive using them in a work environment are different things, and the latter requires study, effort, and above all, perseverance.
Therefore, creating a skills map is also an opportunity to consider your employees’ professional desires and personal motivation, aligning them with the company’s needs, thus creating a beneficial situation for everyone involved. I’m certain that doing this will increase your workforce’s motivation, engagement, turnover, and ultimately, the overall value of your company.
A final word on the importance of storytelling
An extremely crucial and often overlooked skill to possess in the data science world is the ability to tell stories. It’s virtually impossible to have a real impact within an organization if you don’t know how to communicate your results effectively, as no one will understand them or be able to take advantage of their potential.
The main step to effective storytelling is to understand the audience, their background, knowledge, objectives, and roles within the organization. Having clarity on these points helps the presenter to know what information is essential for the audience to understand the goal, the context of the story, and the arguments that explain why the conclusions were reached.
Another factor is to clearly demonstrate the path that has led you to the results you have obtained. Connect the steps you have made during your work, with the business problem you are trying to solve is key to keeping your audience contextualized and motivated during your presentation. And finally, knowing how to describe in a non-technical way the output of your model, including the relationships between the different variables, will help participants understand the overall impact of your results.