spreadsheet-b1+b2
13. November 2018

Spreadsheets are programs 

I’ve said it before, and I’ll say it again: Spreadsheets are programs!

Don’t believe me?

Spreadsheet sample “program” and corresponding Javascript code ↓↓↓

let B1 = 14;
let B2 = 12;
let B3 = B1 + B2;console.log("Total: ", B3);

spreadsheet-b1+b2

let A = [ 10, 13, 8, 2, 7 ];
let B = [ 13, 3, 10, 6, 2 ];
let C = [ ];for (let i=0; i<5; i++) {
 C[i] = A[i] + B[i];
}console.log(C);

sheet-a5+b5

See? Spreadsheets *are* programs. Yet instead of code that is written line by line, spreadsheets allow you to program the relationships between data elements that live in arbitrary cells. Spreadsheet programs also auto update upon every edit, and thereby give the programmer an instant visual feedback loop.

To get theoretical about it: spreadsheets are both functional and reactive ← two of the hottest coding terms of late. (I wrote a bit more about this here)

And there is more. A key idea in computer science is that of three tiered architectures:

  1. data

  2. logic

  3. presentation

The “data layer” in modern web systems typically lives in relational databases.

The “logic layer,” is often a combination of both server-side code and front-end code. The server-side code might be something like Python and the front-end code is typically JavaScript.

Finally, the “presentation layer” — what the person actually sees — is coded in HTML and CSS. Usually there is a host of other technologies involved just to get started.

Enter the spreadsheet. The power — and one of the perils — of the spreadsheet is that it shamelessly mixes the data, logic and presentation layers, like its users couldn’t care less.

Which — frankly — they don’t.

So spreadsheets in general, and Excel in particular, is by far the most used database in the world. It is also the most used programming language and UI tool in the world.

Thus, without knowing it, people that create these types of spreadsheets — and there are hundreds of millions of them in the world — are in fact “full stack” developers.

Back to the spreadsheet vs. coding comparison. The screenshot above shows a fairly typical spreadsheet. It’s a financial model with about a dozen assumptions. Now, if I wanted to recreate this relatively simple spreadsheet using regular programming tools, the logic alone might be 61 lines of non-trivial JavaScript code. Believe me, I tried it:

openspreadsheet

let budget = 10000;
let cpc = 1.0;
let vrr = 10;
let vrl = 0;
let rsr = 2.5;
let rsl = 1;
let churnrate = 10;
let churnlag = 12;
let viralperreg = 1;
let viralpersub = 2;
let price = 19;let series = {
 visitors: [ 0 ],
 organic: [ 0 ],
 paid: [ 0 ],
 registrations: [ 0 ],
 new_subscriptions: [ 0 ],
 churn: [ 0 ],
 registered_users: [ 0 ],
 subscribers: [ 0 ],
 MRR: [ 0 ]
};for (let i = 1; i<37; i++) {
 series.paid.push(
   budget / cpc
 );
 series.organic.push(
   viralperreg * series.registered_users[i - 1] +
   viralpersub * series.subscribers[i - 1]
 );
 series.visitors.push(
   series.paid[i] +
   series.organic[i]
 );
 series.registrations.push(
   series.visitors[Math.max(0, i - vrl)] *
   vrr / 100
 );
 series.new_subscriptions.push(
   series.registrations[Math.max(0, i - rsl)] *
   rsr / 100
 );
 series.churn.push(
   series.subscribers[Math.max(0, i - churnlag)] *
   churnrate / 100
 );
 series.registered_users.push(
   series.registered_users[i - 1] +
   series.registrations[i]
 );
 series.subscribers.push(
   Math.max(0,
     series.subscribers[i - 1] +
     series.new_subscriptions[i] -
     series.churn[i]
   )
 );
 series.MRR.push(
   Math.round(series.subscribers[i] * price)
 );
}

Next, if I wanted it to be presentable on the web, I would need to create a user interface (“presentation layer”) and run a web server.

In order to save edits (like a spreadsheet does naturally) I would have to beef up the “data layer” and run a database system.

All of this would easily take several hundred lines of code in about five different programming languages (e.g. SQL, Python, JavaScript, HTML and CSS).

Say I want to add a chart to visualize the data and I would have to include several different libraries and at least a couple of hundred more lines of code. So you see: spreadsheets are very productive programming environments!

Now, if only there was a way to take existing spreadsheets and easily publish them as beautiful reports, data dashboards or interactive web applications.

Wait. We at GRID are building exactly that!

This article has also been published on Medium.

More blog posts

15. February 2024

Build a website calculator for lead generation in 4 steps

Transform spreadsheets into interactive, lead-generating calculators with GRID, enhancing lead qualification and website conversion rates.

22. November 2023

Tool building in GRID by guest blogger Vincent Doedee

In this guest post by Vincent Doedee, he describes how he's been able to use GRID to leverage his spreadsheets...

27. September 2023

Monetizing expertise through spreadsheet models

The blog discusses monetizing expertise using spreadsheet models, highlighting how domain experts can benefit from GRID's features to enhance their...

22. August 2023

Create beautiful charts with your Airtable bases in 3 simple steps

Learn how to use GRID to make charts and reports from your Airtable bases.

17. August 2023

6 great ways AI can improve your data work

Improve the quality, accuracy, and accessibility of your datavis with GRID’s AI assistants.

08. August 2023

Emre Toker: Empowering entrepreneurs with GRID

How startup founders and investors use GRID to make the best possible decisions.

26. June 2023

How to create a dashboard from your Airtable base

Learn how to visualize your Airtable bases and create a dashboard.

19. May 2023

How to switch from notioncharts.io to GRID: a step-by-step guide

Learn how to easily transition from notioncharts.io to GRID and gain access to powerful features including an AI formula assistant...

17. May 2023

How to create a dashboard in Notion

Learn how to create a dynamic dashboard in Notion using GRID.