Eric Weber

Eric is a passionate data scientist and an inspiration to aspiring and practicing data scientists. Many of his thoughts and advise are 100% applicable in many other areas, including MDM, ERP, CRM and BI.

linkedin.com/in/eric-weber-060397b7

You can subscribe to his newsletter at Substack. I like his thoughts on data as a product.

ericdataproduct.substack.com

Here is a example from May 04, 2022 post on LinkedIn

The get things done mentality is a must - here's why:

1. The data space has no shortage of brilliant people. What is harder to find are those people who are truly able to operate within an organization to get things done. Working with and influencing people is hard stuff.

2. Most of the hard organizational problems aren't purely about data science or machine learning. They are about organizing people and teams around problems that unlock value for the business. That takes persistence, patience and more than technical skill.

3. Building data products is hard because it isn't just about the building, it is about being a product evangelist and motivating people to behave in new ways around data. Creativity is key here.

4. In a new space like data product (at least for most companies), there is a lot of ambiguity in the role and outcome. Ambiguity won't go away quickly and someone's ability to work within it is critical to success.

5. This point is probably more selfish, but I get a lot of energy out of working with those who are unafraid and excited by a tough path forward, full of challenges in user adoption and supporting organizational growth.

Focus on engaging people - not developing more technical skills.

Feeling stuck or unsure what comes next in your data career? Focus on engaging people - not developing more technical skills.

1. Career movement and changes require people, not algorithms. An algorithm can’t tell you what comes next or give you feedback.

2. Talk to people around you. It’s ready isolating to try to solve your career alone. Others have really good input and stories.

3. It’s hard to lean on others when you’re struggling. But if you can start with 1 person to confide in, the relief you feel is palpable.

4. Check out LinkedIn and find some career switchers that show a path you could take. Reach out to them to talk! You’ll be surprised at the response.

5. Spend time doing reflection. A big change doesn’t happen overnight usually. It’s the accumulation of lots of little habits like reflection.

I can’t wait to see what your next step is. Keep moving forward 😊

The basics in data aren't easy.

Fundamentals require consistent practice.

Here are 10 fundamentals I think about all the time:

1. Knowing how to properly use a t-test.
2. Explaining the meaning of a p-value to a non-tech audience.
3. Using WHERE and HAVING properly.
4. Designing a good A/B test so that stats are easy.
5. Defining SQL subqueries in the appropriate cases.
6. Checking the assumptions of linear models.
7. Understanding how to take a simple random sample.
8. Explaining confidence and/or credible intervals.
9. Ability to say in 5 minutes what you could say in 30 min.

10. Never thinking you've got it all down. Review the basics.

    What are your "basics"?

5 things that will make you a better data professional

• at any stage of your career:

1. Ask good questions and improving on them. If you don't define the question clearly, no method or approach has any hope of giving you the results you need, rigorous or not.

2. Determine the quality of the data (or lack thereof) and get better monthly. Understand your lower and upper bound for quality of results. Decide if rigor is even plausible.

3. Decide how exact you need to be. If the decision is high leverage, high impact, being relentless about rigor makes sense. If it isn't, decide if more directional estimates are plausible.

4. Make sure you understand what kind of decision stakeholders need and engage with them about it. Do they need a go/no-go decision, or are they simply trying to learn more?

5. Make clear how rigorous what you've done is and why it was necessary. Simply saying "we used stats" makes everything the same. Help others understand the limitations of your work.

What else would you advise? Any tips?