It was a real help for beginners and students to be oriented and do a first glance at the data and its existing values and distributions per column (see pictures). It also allowed adding new columns with values calculated row-based. I use this for actual data transformation and processing of the raw material while keeping the original.
Will all these features come back, since it helped students and novices in data handling and data viz tremendously? These features were a clear advantage over other tools…
The short answer is no. The longer answer is that there are forthcoming improvements that will provide related functionality with significantly more capabilities.
I discuss the rationale for removing the add cell menu and data table cells here, along with the role of AI for synthesizing common code snippets:
The forthcoming improvements to SQL cells should address much of the functionality previously provided by data table cell, including visual summaries with interactive filtering, but the SQL cell is much more expressive (since you can write SQL) and therefore less of a “dead end”.
But in terms of systematics, it shouldn’t be a data table cell the parent/generalization of an SQL cell so that all analytical features are available to all other children (CSV, SQL, etc.). The data table provided very easy access for people with early coding skills (as novice students are).
It seems that easy access features, also like the “add cell menu,” are no longer the focus, which made it attractive for choosing it for the lab class of my InfoVis course where the students liked it very much for three years now working with the old observable interface, especially since it had a good balance between "visual support, " including telling you its capabilities visually up front, and focussing the programming on the visualizations themselves. They really enjoyed the lab classes with Plot in the old notebook, and I even used it for the exam.
Wasn’t it the idea to help people with “limited/early” coding skills to focus on creating visualizations with the reduced OJS and visually/interactively supporting the rest? I can tell from my course that this worked.
I’m sorry about my tone, but that is a bad surprise given that my course is only 4 weeks away.
You’ll be able to use a SQL cell to query CSV and other tabular data, so my hope is this will be similarly convenient to get started but without the severe functional limitations of the data table cell (such as the inability to join tables).
We’re still very much interested in making data analysis and computation more generally more accessible and interpretable. But we’re trying to avoid “dead ends” that are easy to get started but do not scale to meet real-world needs. We’re hoping that our new approaches will offer significantly more capabilities and help people go much farther than they could before.
Of course, the proof of the pudding is in the eating, so we please wait to pass judgment until the new features are available. Until then (and for some time beyond), please continue to use Old Observable if you prefer it.
