Spreadsheetalikes

October 15, 2014 — August 2, 2022

computers are awful
faster pussycat
number crunching
office
UI
Figure 1: Spreadsheet from 1700 BCE, according to Daniel Mansfield and Norman Wildberger

Excel: R or jupyter for people who are prepared to pay for an inferior version of a free app.

In my view, the unique selling point of spreadsheet is providing a GUI for your data. This is not much of a selling point since a non-spreadsheet statistics programs like R can also have a GUI and better graphing and analytics besides, and the spreadsheet GUI is obtrusive and limiting by comparison.

I say this as someone who finds the R language design icky. Spreadsheets are worse. They are good at viewing certain types of data but awful at non-trivial things. For a wide variety of reasons, they are used for non-trivial things. See Packy McCormick, Excel Never Dies:

There are an estimated 1.2 billion people who use Microsoft Office, and while it’s hard to know exactly how many people use Excel regularly, estimates put it at 750 million users. By comparison, as of 2018, there were only 10.7 million Javascript developers and 7 million Python developers.

But this GUI business does pays for itself when it comes to collaborative editing or analysis of your data. In the 21st century there are online collaborative spreadsheets. Sharing the workload is a good thing; hardly a data set worth having was collected by one person. (Although I do want a sneakernet version of various of the online apps).

It can be weirdly beautiful though: Top Excel experts will battle it out in an esports-like competition this weekend.

Anyway, sometimes we need spreadsheets for reasons of compatibility or because the task is actually naturally a spreadsheet-y task. What do we do then?

1 Spreadsheet interfaces to other tech

See Lucy Huggins’ defence of the affordances of spreadsheet UI for data entry, based on annotating historical free-form data.

1.1 Textual tabular formats

Just want to get at some tabular data as plain terxt? See Text data processing.

1.2 Jamovi

Jamovi

jamovi is a new “3rd generation” statistical spreadsheet. designed from the ground up to be easy to use, jamovi is a compelling alternative to costly statistical products such as SPSS and SAS.

jamovi is built on top of the R statistical language, giving you access to the best the statistics community has to offer. would you like the R code for your analyses? jamovi can provide that too.

jamovi will always be free and open - that’s one of our core values - because jamovi is made by the scientific community, for the scientific community.

The interface looks good. It is lacking certain features that I would like (generalized linear models, AFAICT interaction terms in regressions) but it does good testing and analysis, and nudges you towards best practice especially in hypothesis tests.

You can develop extra extensions.

1.3 Userland

userland is a hybrid of spreadsheets and dataflow patching.

1.4 Honeycode

honeycode is a hosted service from Amazon that allows you to use a spreadsheet-style interface to generated real codeless apps. Maybe that makes it Airtable with big business backing, or google sheets with an app-platform-first orientation.

1.5 Streamsheets

Streamsheets is a spreadsheet for real-time live-updating data. They seem to be node-based, FWIW, and open source. (source).

2 Web-augmented classic spreadsheets

Here is a list of online collaborative data-analysis spreadsheet-like tools, which justify the awkwardness of a spreadsheet-like interface by being shared.

  • Luckysheet embeds a spreadsheet element in a webpage

  • jExcel is not a full spreadsheet implementation but a javascript widget for implementing spreadsheet interfaces for online stuff. Even handles spreadsheet formulae.

  • ethersheet is a node.js-backed cypherpunkish spreadsheet made by privacy warriors, much like sibling project etherpad. Requires MySQL.

  • ethercalc is another node.js-backed spreadsheet with

    • explicit design documentation,
    • low-fuss backends (redis or fs)
    • and an explicit pedigree starting with Dan Bricklin, the progenitor of all spreadsheets.

    On the other hand,

    • It may be a one-person shop, and
    • the interface is dilapidated vintage classic.

    Kind of a HD text mode, with certain shortcomings such as lack of column selection.

  • Google offers a spreadsheet, although I do try to avoid their products.

  • airtable is another unorthodox productivity/managing/note-taking thing which I had recommended to me but have not investigated yet. It seems to be a full-featured app platform.

3 Probabilistic spreadsheets

If you can’t estimate the required quantities from data, can you improve the estimates using quantified guesses? Yes, this is a well-studied phenomenon.

3.1 Guesstimate

Guesstimate is a neat proof-of-concept probabilistic spreadsheet, which does Monte Carlo propagation-of-error. (source, Blog post).

For a good worked example, see Elizabeth’s blog.

3.2 Enterprise probabilistic spreadsheets

Enterprise versions of guesstimate.

4 New twists on classic sheets

  • A lot of the tasks that end up as spreadsheets would be better as jupyter notebooks.
  • treesheets is an open-source data GUI which is not really calculation-focussed but more a hierarchical freeform database.
  • Soulver is a macOS/iOS app that will do plain-ish language calculations with smart currency conversion etc. (USD30)
  • Microsoft Power Fx overview

5 Interfacing with spreadsheets

6 google sheets hacks

You can send email from google sheets, fancily.