# ISN Lab: Research you can look at

> An interactive research portal for SISJ. Researchers' own CSV, Excel, and PDF files become maps and charts on the page.

- Site: Falah Isnan (https://falah-isnan.com/work/isn-lab)
- Author: Falah Isnan (https://falah-isnan.com/)
- Published: 2026-09-23
- Updated: 2026-09-23
- Discipline: Product engineering / data visualisation
- Focus: Flexible intake, honest output
- Year: 2021
- Role: Feature engineering — the data-to-visual pipeline
- Client: SISJ — Society of Indonesian Science Journalists
- Stack: Laravel / Next.js / Tailwind CSS
- Access: Live site

## The problem

Every researcher submits a different format. A portal that only accepts one shape of data is a portal that publishes summaries of research instead of the research itself.

## The approach

Read the file first — CSV, Excel, or PDF — work out what is actually in it, and only then choose a visualisation the data can honestly carry.

ISN Lab publishes research on the web the way research is actually read: by looking at it. A Laravel backend holds the material, a Next.js and Tailwind front end presents it.

My contribution was the part in between — the feature that takes a researcher's own files and turns them into interactive maps, charts, and other visualisations on the page.

## What the project is

### A portal for research, not for papers.

The point of the portal is that a finding can be explored rather than described. A reader should be able to move through the data a researcher collected, not read a sentence summarising it.

That only works if the visualisation comes from the researcher's real data. Retyping numbers into a chart tool would have made the portal a publishing surface instead of a research one.

### Every researcher sends a different shape.

The files arrive as CSV, as Excel, and as PDF. None of the three agree on where a header sits, how a region is named, or whether a column holds a number or a note about a number.

There was no format to standardise on, because the material was already collected. The intake had to accept what researchers actually had.

### Read the data before choosing the visual.

So the tool reads the file first and works out what is in it — which columns carry places, which carry series, which are labels — and only then decides what can honestly be drawn from it.

That order matters. A map is offered when there is geography to put on it, a chart when there is a series to plot. The data decides the visualisation, not a template.

## The hard part: Formats that never agree

The hard part was not drawing the chart. It was deciding what the file in front of you could support.

1. **Take the file as it is.** CSV, Excel, or PDF, with whatever layout the researcher used. No preparation step, no template to fill in first — the intake starts from the file that already exists.
2. **Work out what is actually in it.** Find the columns and read what they hold: place names, numeric series, dates, labels. The result is a description of the data, not an assumption about it.
3. **Choose a visual the data can carry.** Geography becomes a map; a series becomes a chart; anything the file cannot support is not offered. The flexibility is in recognising the data, so the output can stay honest.

## What you can open

- [SISJ Indonesia](https://sisj-indonesia.org/)
- [ISN Lab](https://www.isn-lab.com/)