Chapter 16: Conclusion and Outlook
What Is Data Science (Again)?
- Data science encompasses:
- Detection
- Information
- Computing technology
- E-marketing
- Social media promotion
- Services
- Models communication
- Multimedia
- Processing
- Computer
- Network projects
- Predictive consumer organization
- Branding
- Consumer demand markets
- Planning web marketing
- Coding
- Program analytics programming
- Events
- Data mining
- Machine learning vision engineering research reality
- Solutions
- Web services
- Maths pattern
- Engineering
- Planning media
- Statistics
- Big strategy worldwide
- Web dev KDD
- Data visualization
- Mobile pricing
- Probability
- Computing
- Segmentation
- Target digital
- Social networks
Defining Data Science
- Data science is still evolving as a well-defined field.
- There isn't a standard, universally accepted collection of knowledge.
- Significant overlap exists with other disciplines.
- The field is subject to hype and exaggeration.
- A central authority to correct misinformation is lacking.
- This ambiguity makes curriculum development challenging.
Learning Data Science
- Some industry professionals argue that data science is best learned on the job, rather than in academic settings or through books.
- However, the lecturer believes that there are fundamental principles that can be taught.
The Art of Data Science
- Many practitioners jump directly to complex algorithms.
- However, there are crucial steps and considerations before that.
- One vital aspect is the infrastructure required to support these algorithms.
- Functionality can be broken down into three areas:
- Data management
- Data processing
- Resource management
Industry Insights from ZHAW Survey (2022)
- Survey by Mildenberger & Stockinger at ZHAW in 2022 revealed:
- Machine learning is gaining traction in industry.
- Data engineering skills, particularly in data management (e.g., SQL & DWH), are highly sought after.
A Word of Warning: Potential Pitfalls
- Data science systems can handle vast amounts of data, enabling analysis and model building.
- However, these models can inadvertently encode prejudices, misunderstandings, and biases.
- Increasingly, these models govern aspects of our lives.
Addressing the Pitfalls
- Cathy O’Neil refers to problematic models as “Weapons of Math Destruction.”
- The systems and data aren't going away.
- Actions to mitigate the negative impacts:
- Increase transparency
- Implement checks and balances
- Ensure accountability for responsible parties