DUOLABS
Work

Statistical weighting visualization tool

An internal statistics chart tool for a national research institute, where researchers assign weights to several data items and compare the results as charts. The tool was only useful if every change in weights was recalculated and shown on the chart right away. We built the weighted chart screens, the logic that applies and recalculates weights, and how the charts are displayed and operated.

Launch
2009
Client
A national research institute
Type
Data & Visualization
01 / 04

A chart that redraws in place when a weight changes

The weighted statistics chart aggregates research data according to the weights a researcher sets. We built the calculation logic and the screen together so that when a weight changes, the result is recalculated and reflected on the chart right away. Only when the change and its effect sit on the same screen can a researcher tell which weight moved the result.

A chart that redraws in place when a weight changes
02 / 04

Weights assigned item by item

The weighted statistics tool produces its result by assigning weights not to one data item but to several. We implemented the weighting logic so that each item gets its own weight and the value set goes into the calculation exactly as entered. A researcher can interpret the result only when it reads together with which item was given what weight.

Weights assigned item by item
03 / 04

Comparing how results shift with the weights

The purpose of the weighted statistics tool is not to show one right answer but to let researchers compare and analyze how results change with the weights. We set up the chart display and controls so researchers could compare and analyze results. The same data ranks and scales differently depending on the weights, and reading that difference is the researcher's job.

Comparing how results shift with the weights
04 / 04

A display structure that lets numbers read as meaning

When statistical results are just rows of numbers, the researcher has to draw the picture again in their head. We organized how the charts display and respond, refining the visualization so numeric data reads meaningfully on screen. The standard for the layout was that results follow visibly from the values the researcher set.

A display structure that lets numbers read as meaning

Highlights

01

A statistics chart that aggregates several data items by assigned weights

02

Charts that recalculate and redraw as soon as a weight changes

03

Chart display and controls built for researchers to compare and analyze results

Insights

Read the blog