Keeping Up With Data — Week 38 Reading List
Data, data, data. Everywhere we look, right? And the trend is accelerating as we can see in the image above. It is worth opening the infographic — as it includes charts speaking to all sorts of aspects of data: Is technology a bigger issue than people/culture when becoming data-driven? What are the most commonly reported benefits of using data effectively? Or what share of front-line workers don’t have access to data and analytics?
The upward trend of data has also been demonstrated by a massive Big Data LDN conference — with a ‘join the data revolution’ tagline — taking place this week.
A trio of articles speaking to strategic, tactical, and operational elements of data is on the menu today.
- The future of work will likely be filled with human-machine partnership: Data, as the image above indicates, is growing. And the importance of leveraging data for business benefits is gaining momentum. But while the (data) technology is often clearly immature, the main blocker — many conclude — is in us and our ability to ‘speak data’, be data driven and AI literate. The future requires our “ability to engage with and take advantage of [data] needs to increase and adapt in step”. A great article that speaks my mind when it comes to data literacy and the bridges between data science and business. Though I wouldn’t necessarily go as far as requiring everyone to “have at least a fundamental understanding of ML, NLP, robotics, and deep learning.” Or not yet. Understanding the data as a mirror of the world seems ambitious enough to me now. (Mail & Guardian)
- The Leader’s Guide to Being Data-Driven in 2021 (Part 1): Being data-driven comes with many benefits. But becoming data-driven often comes along with meaningful costs. But does it have to? Can’t we leverage the existing data investments better? Building on her experience from Netflix, Michelle offers three tips: cultivate a “one team” mindset; pick an audacious goal and establishing a “Data Success” team; and establish success-focused standards and evangelised consistency. The first is important because it literally takes everyone to be data-driven (from front-line to back-end engineers to execs). Next you create a task force of your top performers from all functions and assign them a big hairy audacious goal. Once they succeed, you make it a standard. Because “the greater the consistency, the better the scalability”. Worth giving a shot, don’t you think? (Michelle Ufford @ Noteable)
- Phik (𝜙k) — get familiar with the latest correlation coefficient: The last article speaks about the correlations. Pearson’s r is often used to measure a linear relationship between two continuous variables. When categorical variables are involved, Cramér’s V is considered. The latest correlation coefficient — Phik (or 𝜙k) — tries to make the correlation analysis more convenient by working consistently across categorical, ordinal, and numerical variables, being less sensitive to outliers, and capturing also non-linear dependencies. Sounds great! Of course, it’s ‘no free lunch’ as the calculation of 𝜙k is computationally heavy, less precise for continuous variables and provides no indication of the direction of the relationship. Either way, pleasure to meet you, 𝜙k! (Eryk Lewinson @ TDS)
I’ll be giving a talk titled: “Data literacy: symptom or cause?” next week, so a lot of reading was done as part of researching the topic to help organise and test my thinking.
Another big topic for me was visualisation as it was one of this week’s modules of Data Science for Business course I’m taking at Harvard Business School Online. Never stop learning, right?