Role Basics: What is Data Science?
Data science involves anything from analyzing detailed operations or marketing data to building machine learning or predictive models. You spend a lot of time with data, and writing code.
Flavors of Data Science roles:
Organizations vary in how technical their data science roles are: some might require years of specialization (equivalent to an advanced degree), while others are accessible within a few months (similar to what you’d get from a “data science bootcamp”). In addition, organizations may apply data science to a subset of three broad domains: operations, marketing, or the product itself.
Example projects in Data Science:
- Predict which suggested products will optimize customer conversion: Use customer data to improve the recommendation algorithm (which could be recommendation of a product to try, of content to read or anything else that increases customer value).
- For a marketing automation company: Identify which email messages are likely to increase conversion outcomes.
- Use predictive modeling to improve sales: Predict which customers are likely to churn soon, and which are likely to be receptive to upselling, so we can allocate sales and customer success resources more effectively.
- Predict which clients are most likely to default on a loan.
- Use inventory, transaction and other data to optimize inventory management.
Common activities in Data Science:
- Work with business stakeholders to analyze data using tools like excel, R, and python.
- Build algorithms and predictive models using R or python.
- Use SQL to query the database and construct data sets.
- Work with engineering on implementing back-end features based on successful projects.
- Create visualizations that help explain data analysis findings to non-technical business stakeholders.
Data Science metrics:
This depends a lot on what departments you’re working in and will typically roll up to the teams’ metrics, including things the team is trying to optimize like:
- Inventory levels
- Product engagement
- Customer churn rate
- Customer lifetime value
- Customer conversion rate
- Cost per acquired customer
Data Science Compensation:
Beginning jobs often pay $100K or more, though these often assume several years of relevant experience.
Data Science career path:
This can be a great career option. First, it’s fantastic for individual contributors, as predictive modeling and data analysis can be applied in many fields — from any large tech company, to retail, to advertising, to finance. You can be a highly paid contributor without ever needing to manage people. But you can also use data science to branch into business and management roles like growth, product management, and more quantitative marketing roles.
How accessible are Data Science jobs?
Low. This is a harder role, because it requires both talent and interest in data analysis, and some training. It is, however, relatively transparent — great for autodidacts.
- Time to learn. 1-2 years from scratch, or 3-9 months assuming the equivalent of several college level quantitative classes.
- Selectivity. Highly selective.
- Ease of working remote. Generally, remote-favorable.
Job Requirements: What you need to be competitive for data science roles?
Key skills for Data Science:
- Excel
- SQL / Database
- R or Python for complex data analysis, algorithm development and modeling
- Written and verbal communication
Professional background for Data Science:
This role doesn’t require any particular background if you have the skills. People often develop the appropriate experience through:
- Undergraduate or graduate study (typically in math, computer science, physics, engineering, economics or another highly quantitative field).
- Self-instruction
- Work in a data-intensive field like advertising or operations, where you end up really enjoying the analysis and go “down the rabbit hole” of increasingly technical projects:
- Conduct analysis of product or operations that significantly improves the business (i.e. 10% improvement of a key metric or pays for your salary 5X ore more over that period).
- Built a model or algorithm that is integrated into company operation models.
- Completed an interesting side project with obvious applications.
Personal characteristics for success in Data Science:
- Very quantitative
- Tolerant of spending hours or days
- Curious
- Happy focusing on work and not interacting with people
How to prepare for and get a job in Data Science:?
Data Science projects to learn and prove yourself:
- Kaggle projects
- Netflix prize data set
- Take on responsibility for A/B testing and site optimization for a business with web traffic of 10K per month or more
- Build a predictive model that lets you trade effectively on high-velocity PredictIt markets
- Use data from Twitter to identify high-value prospects for a product or service that a company or organization can reach out to directly
- Implement a chatbot for a person or organization that has a lot of Q&A data (like an online course with 1000s of past students)
- Work as an RA for a business school researcher who is doing large-scale data analysis